docs: seal reading reliability S release and next live validation plan

This commit is contained in:
Agent Workbench Dev
2026-11-04 01:35:36 +08:00
parent 5dcd6d7c3f
commit 8f41611b1b
778 changed files with 1646308 additions and 4 deletions
+7
View File
@@ -226,3 +226,10 @@ docs/live-delivery-r-evidence/**/*.log binary
docs/live-delivery-r-evidence/**/*.txt binary
docs/live-delivery-r-offline.txt binary
docs/live-delivery-r-git-index.txt binary
docs/reading-s-evidence/** -text -whitespace
docs/reading-s-evidence/**/*.log binary
docs/reading-s-evidence/**/*.txt binary
docs/reading-s-offline.txt binary
docs/reading-s-git-index.txt binary
docs/reading-s-release-identity.json -text
+2
View File
@@ -43,3 +43,5 @@
!/docs/network-q-evidence/**/*.log
!/docs/live-delivery-r-evidence/**/*.log
!/docs/reading-s-evidence/**/*.log
+3 -1
View File
@@ -1,5 +1,7 @@
# Agent Workbench项目执行规范
S已本地交付1.24.7/source5dcd6d7c/候选20261103164618190:文件续读携带原范围/偏移/SHA、纯文件批次容量及未用份额分配、Run/Task预留余额准确解释。96相关和18安装升级、源码/便携/实际安装新UI各10组、20组历史UI、565输入独立重建22/22、三归档/便携及12实际安装和16+9文件流程通过。正式原PID36552/Provider、63项目79会话6262记录13279事件及outbox/3660工件,所有原表行/全导出/工件与公开模型中心在网页/原生回看后再次精确保持,9229关闭。760份/106829868字节证据离线通过,见docs/QA-READING-RELIABILITY-S.md及docs/reading-s-release-identity.json。原R三响应+固定续读回放由8轮未完到6轮读完2584;原九测试仍失败,0修改/交付,本轮真实模型请求0。下一按docs/LIVE-DELIVERY-T-PLAN.md另冻结单次预检/新Node,16K/2048/8请求/90000tokens;不续跑旧失败,Node通过再React/Python/Purge。更新869425ms重复核验及1.25独立排期。不委派、不提取/复制/重存Key、不重启原Provider、不推送。
R已完成Q后的新有界真实样本:唯一所属诊断DNS/auto/IPv4/IPv6 TLS通过(1579ms,0HTTP/模型/凭据);新预检1请求/4492报告tokens/0工具/strict matched。新Node6请求/72571报告tokens/0未知、全JSON matched,12工具/1原检查审批,原9测试0通过9失败/基线build通过,0提案/修改/交付;第7次本地输入12352可容纳,但预留29623>剩余17429而token_limit暂停,非新context_limit/网络失败,不续跑。原61项目77导出/3610工件/209尝试和原行(项目位置自有前插2逆算)保持,当前63项目79会话6262记录13279事件及outbox/3660工件,新增7尝试全自有。预算/高性能池恢复,目标5→6→7/凭据1、全局/legacy65→69,9229关闭,正式仍1.24.6/source54dbe3d6/原PID36552/Provider。61自有工件记录全取回,178份/75944723字节证据离线通过,见docs/QA-LIVE-DELIVERY-R.md。下一优先docs/READING-RELIABILITY-S-PLAN.md:离线复现文件续读定位、约400字符分页吞吐及预留暂停解释;不移除预算保护或篡改上游27164输入报告。先修复/独立验收,再另立新Node;React/Python/Purge收益和1.25仍按前置计划。不委派、不提取/复制/重存Key、不重启原Provider、不推送。
Q已本地交付:1.24.6/source54dbe3d6/候选20261103151145294,所属DNS/TCP/TLS有界诊断,0凭据读取/HTTP/模型调用。69相关、11宿主、14安装旧版升级、4补强热升级、新UI源码/便携/实际安装各9组、五类历史26组、560输入独立重建22/22、三归档/便携、12实际安装和16+9文件流程通过;其中555跟踪输入及5例示dist均独立核对。正式保留原PID36552/Provider/61项目77会话/6167记录/13115事件及outbox,全部原表行/导出/工件和公开模型中心在网页/原生回看后再次精确保持,9229关闭。438份/89228582字节证据封存离线通过,见docs/QA-NETWORK-DIAGNOSTICS-Q.md及docs/network-q-release-identity.json。实际更新677755ms重复核验单列性能。下一新有界预检/Node,不重放旧失败;React/Python/Purge收益和1.25仍按计划。不委派、不提取/复制/重存Key、不推送。
@@ -122,7 +124,7 @@ D固定可信来源生产策略已完成:实现9106655/最终验收d05a741,6
## 当前本地服务与Windows
- 正式服务入口http://127.0.0.1:4318,当前1.24.6,沿用原PID36552与Provider;不可变程序身份以docs/network-q-release-identity.json为准,R后配置/数据与实测见docs/QA-LIVE-DELIVERY-R.md。R已按声明核对原全表/导出/工件保留及原预算/池恢复;普通源码/规划/Git工作不重启服务、不读取或复制Key。正式原生窗口按Q实际回看记录,R没有新增原生/UI验收结论,旧F的Escape失败作为历史保留。
- 正式服务入口http://127.0.0.1:4318,当前1.24.7,沿用原PID36552与Provider;不可变程序身份以docs/reading-s-release-identity.json为准,S正式保留状态与网页/原生回看见docs/QA-READING-RELIABILITY-S.md,最近真实模型仍是R未交付原样本。S核对原全表/导出/工件/模型中心精确保持,普通源码/规划/Git工作不重启服务、不读取或复制Key。下一按docs/LIVE-DELIVERY-T-PLAN.md另冻结新样本,旧F的Escape失败保持历史含义。
- 浏览器/fixture验证使用独立端口与数据目录;用户项目与研究对象的登录/会话配置不用于测试凭据。
- Windows执行环境以实测shell为准。现有工具使用PowerShell 5.1和UTF-8;不用&&/||拼接命令,使用换行和$LASTEXITCODE核对结果。
- 桌面与移动共用执行宿主。手机127.0.0.1指手机,远程接入通过未来经验证的配对网关。
+1 -1
View File
@@ -1,6 +1,6 @@
# Agent Workbench 更新记录
## 1.24.7 · 文件续读与预算解释(发行验收中)
## 1.24.7 · 文件续读与预算解释(已本地交付)
- 文件续读入口携带原行范围、偏移和可选 SHA-256 核对;读取版本变化会拒绝拼接,已返回页工件与未读原文件后文有明确区别。
- 纯文件读取依据实际可用消息容量分配,同批小文件用剩的空间可供后续读取;紧张容量下优先保证完整页能进入下一请求。混合操作和工件读取保留原保守策略。
+19
View File
@@ -0,0 +1,19 @@
# T:读取修复交付后的新真实 Node 样本
本计划接续 S 1.24.7 的本地发行。S 只做隔离回放和可靠性修复,未请求真实模型,也未证明自主交付或 Purge 收益。T 的“新样本”需另行冻结;本文件不能作为旧 Task 自动续跑指令。
## 执行顺序
1. 冻结新的正式版本/候选/source、原 PID/Provider、当前高性能池、目标连接公开修订/凭据修订和全部原表行、导出、工件摘要。当前启用的 jojocode GPT Chat JSON 是预选,执行前重新读取实际状态;不复用过期修订,不提取、复制或重存 Key,不自动启用旧 apikey。
2. 为所属连接按原 Q 范围作一次有界 DNS/TCP/TLS 观察。它不发 HTTP、模型请求或读取 Key;不改变系统代理、IP 族、证书校验或健康评分,不能将握手通过当成 API 稳定性证明。诊断异常按所属结果结束。
3. 另开无工具单次身份预检。固定同一连接/协议/端点/严格型号策略及新请求预算;上游原文、实际 URL、型号和用量均封存。无重试、切池或别名补造;失败到此结束。通过仅证明这个样本的协议与报告身份,不认证底层权重。
4. 新建隔离 Node 费用报表工程和新 Task。原三源码/九测试、原目标及约束字节保持;16K 上下文、2048 输出、1600 安全预留、最多 8 请求/90000 tokens/600000ms,单条已通过预检的连接。原用量预留、身份保护、停止和审批保留。
5. 只由模型完成定位、实现三个模块、完整变更集、修复及交付计划。监督只核对/批准其原计划检查、审查应用完整三文件变更集及确认本地 Git;不提供源码、修复答案或补充提示,不修改九测试、配置、owner 文件和原暂存,不增加预算,不使用 run_command。
6. 分别核对每次实际请求的文件新覆盖、完整消息预算、已保留源码/约束、原上游报告用量及 Tool/Task/Run 所属。发生暂停、网络错误、未知用量或身份拒绝时封存原结果,不自动重放或续跑。
7. 成功需要原九测试和 build 通过、当前源 SHA/检查/变更集及 Git 计划一致,最终仅三个允许源码的已确认本地提交;口头声明、读完文件或基线 build 通过不算交付。结束后恢复本次允许暂改的公开预算/池/身份策略,按精确声明解释修订单调增长,逐原表行/导出/工件和公开模型中心核对,9229 关闭。
## 后续门槛
Node 通过后,再按独立冻结预算扩展 React 页面/交互和 Python 原测试交付。原版 Purge AB/BA 必须另设样本,锁定 Harness/Purge/profile、同工程/模型/预算、提示和工具/源码输入,保存失败及全部已报告/未知用量,分别评价交付、步骤/请求数、时间和成本;不把插件加载、提示增强或单次成功称为收益。
1.25 的任务要求/交付复核、其余能力和更新的重复完整目录核验性能保持独立排期。性能优化先测重复遍历的调用来源及不变身份缓存边界,保留字节/目录/指针、活动窗口/请求保护、故障回退和原 Provider/PID;不得为加速删核验。
+4
View File
@@ -1,5 +1,9 @@
# 后续执行计划
S已本地交付1.24.7/source5dcd6d7c/候选20261103164618190:文件续读携带原范围/偏移/SHA、纯文件批次容量及未用份额分配、Run/Task预留余额准确解释。96相关和18安装升级、源码/便携/实际安装新UI各10组、20组历史UI、565输入独立重建22/22、三归档/便携及12实际安装和16+9文件流程通过。正式原PID36552/Provider、63项目79会话6262记录13279事件及outbox/3660工件,所有原表行/全导出/工件与公开模型中心在网页/原生回看后再次精确保持,9229关闭。760份/106829868字节证据离线通过,见docs/QA-READING-RELIABILITY-S.md及docs/reading-s-release-identity.json。原R三响应+固定续读回放由8轮未完到6轮读完2584;原九测试仍失败,0修改/交付,本轮真实模型请求0。下一按docs/LIVE-DELIVERY-T-PLAN.md另冻结单次预检/新Node,16K/2048/8请求/90000tokens;不续跑旧失败,Node通过再React/Python/Purge。更新869425ms重复核验及1.25独立排期。不委派、不提取/复制/重存Key、不重启原Provider、不推送。
S已完成原计划。下一优先[T新真实样本](LIVE-DELIVERY-T-PLAN.md),先单次预检、同预算新Node,再依前置条件扩React/Python及原版PurgeAB/BA。下方R及更早段落按其历史时点理解。
R当前:Q已交付1.24.6的所属连接诊断;后续唯一诊断DNS/auto/IPv4/IPv6 TLS通过,新单次GPT预检1请求/4492报告tokens通过。新Node6请求/72571报告tokens/0未知,全原JSON matched,原9测试0通过9失败、基线build通过,尚无修改/交付。第7次本地输入可容纳,因保守预留29623>剩余17429发送前token_limit暂停。原61项目77导出/3610工件/209尝试和原行、预算/用户高性能池保持,178份证据与Git原字节索引核对通过。见[结果](QA-LIVE-DELIVERY-R.md)及[回顾](RETROSPECTIVE-LIVE-DELIVERY-R.md)。
下一先按[读取可靠性S计划](READING-RELIABILITY-S-PLAN.md)离线复现并改善文件续读定位、实际分页吞吐和预算暂停解释,独立产品验收后另开新有界Node。保持原16K/2048/8请求/90000tokens及安全预留/上游原报告,不续跑旧Task;Node通过再React/Python和原版PurgeAB/BA。1.25其余能力和更新重复目录核验性能保持独立排期。下方P/O/J等“最新”段落只代表当时历史时点。
+31
View File
@@ -0,0 +1,31 @@
# S:读取可靠性与预算暂停解释验收
1.24.7 已本地交付,source `5dcd6d7c3f4f465071ae1156bb1e0bf5a28f48e2`,不可变候选 `Agent-Workbench-1.24.7-win-x64-20261103164618190`。固定发行身份见 [reading-s-release-identity.json](reading-s-release-identity.json)。正式交付完成时间以原 [最终收据](reading-s-evidence/receipts/formal-release/final.json) 为准;包内“发行验收中”文档是冻结时点快照,未为补写结论改包。
## 本次行为
文件续读明确区分恢复已返回页与读取未返回的原文件,携带原路径、行范围、偏移和可选 SHA-256;版本变化拒绝拼接。纯文件批次按完整转义消息预算分配并回收同批未用空间,混合操作与工件读取保留原保守预留。旧发送视图保持可核验。下一请求预留超出 Run/Task 余额时,分别展示所需与剩余;原预算、1600 安全预留、目标/约束、观察用量和上游原文保持。详见 [实现说明](READING-RELIABILITY-S-IMPLEMENTATION.md)。
## 已完成验证
- 96/96 相关测试及 18/18 安装/旧版升级回归通过;TypeScript/Vite、移动、服务与桌面构建通过。
- 源码、便携和实际安装副本的新暂停界面各 10 组通过;1440px/390px 四类诊断、刷新、续接预览及取消保持原 Task/Run,未发出模型请求。历史上下文 11、预算 2、模型选择 7 组通过;390px 是浏览器视口,本轮未重新做 iPhone 实机联调。
- 三协议实际请求验证 Unicode/emoji、CRLF、长行和小/大源码页、历史版本及实际发送容量。R 原前三响应加固定续读策略:旧实现 8 请求仍只在后续请求可见到 2173/2584,新实现 6 请求完整读取 2584。原九测试仍全部失败、build 通过,0 源码修改或 Git 交付;这是离线夹具改善,不是真实模型成功或 Purge 收益。
- 源码 ZIP 565 输入独立离线重建,22/22 编译输出与不可变便携一致;560 跟踪输入逐 SHA 对照,另 5 份示例 dist 用原 ZIP 源码/脚本独立重建匹配。三归档 CRC、清单、目录、字节与 SHA 核对通过;实际便携启动、REA/受控引擎/浏览流程通过。
- 12 组实际隔离安装、窗口/请求/审批保护、清单故障、指针写入后回退、正常更新、原 PID/凭据保留、重启和卸载通过;实际安装副本的成果 16 组、文件/HTML 9 组及 S 新 UI 10 组通过。测试只用自有本机 HTTP 和凭据夹具。
## 正式升级与回看
实际 Setup 将 1.24.6 升级到 1.24.7,保留原 PID36552 和 Provider。63 项目、79 会话、6262 记录、13279 事件及 outbox、3660 工件文件的全部原表行、全导出、原工件字节和公开模型中心精确保持;默认高性能池、三条原连接/原修订和旧 apikey 停用状态保持。没有读取、复制或重存 Key,没有重启原 Provider,没有恢复旧 Task。
正式网页 1440px/390px 阻断全部 API 写入,实际旧 R 暂停显示输入 12352 可容纳、预留 29623、余额 17429,未改写历史或续跑。正式原生窗口经原安装入口打开后回看;其后再次核对全部原状态,9229 关闭。原生原图格式与文件扩展名已核对。
本次安装操作 869425 ms,驱动总耗时 888315 ms。完整清单与目录重复核验保持,详见 [耗时](reading-s-evidence/receipts/release/update-timing.json),后续性能优化独立验证。
## 原失败与证据
首计划 ID/混合视图判据、读取分配与历史版本回归失败、UI 用量字段错误及修正全部保留。首便携启动被 HOST_PORT_UNVERIFIED 拦下;确认隔离目录无启动产物、无所属宿主后,另建隔离样本完整通过,未改候选或降低保护,唯一原因仍未知。取证助手首清单字段错误、正式网页验收器旧定位/隐藏栏/弹窗加载等待失败,以及原生最小化/用户输入拒绝和启动窗口迟到同样保留;后续修正定位/等待并用真实回看确认,没有重装或改写正式记录。
[证据清单](reading-s-evidence/manifest.json)含 760 份、106829868 字节,SHA-256 `612aacef74d95b5158f2bf1b64f584a26829819fe525501436657a42fd5eb767`。新旧源码、原 R 输入引用、请求/计划回放、首失败、最终测试、独立源码/发行/安装、正式状态和原图按原字节封存;旧源码 node_modules 连接未递归复制。
[离线核验器](../scripts/verify-reading-s-evidence.mjs)逐文件及语义核对通过,见 [收据](reading-s-offline.txt)。加 --index 核对 Git 暂存原字节,收据为 [reading-s-git-index.txt](reading-s-git-index.txt)。本地范围提交,不推送。阶段取舍见 [回顾](RETROSPECTIVE-READING-S.md),下一按 [T 新真实样本计划](LIVE-DELIVERY-T-PLAN.md) 执行;本轮真实模型请求 0。
+1 -1
View File
@@ -1,6 +1,6 @@
# S:文件续读与预算暂停修复
本切片接续 R 原失败,产品版本 1.24.7,发行验收中。正式升级前仍以 1.24.6 为准。后续发行身份、保留状态与原始检查结果另见最终 QA,不用开发回放代替上线或真实模型交付。
本切片接续 R 原失败,1.24.7 已完成本地发行及保留原 PID/Provider 升级。发行身份、状态保留与原始结果见 [最终 QA](QA-READING-RELIABILITY-S.md);开发回放与真实模型交付分别判定,候选包内该文档保留冻结时点原文。
## 文件读取
+27
View File
@@ -0,0 +1,27 @@
# S 回顾:读得到、带得走、解释得清楚
R 暴露了三个可定位的问题:原文件后文和原工具页包装的入口混淆;16K 完整请求的固定开销使分页落到约四百字符;下一请求观察用量预留超出余额,却被笼统解释成目标或上下文过大。S 围绕这三个问题修复,保留原模型报告和预算保护。
## 采用的实现
续读来自原页的路径、行范围、SHA 和下一偏移,读取时可核对原字节。原范围完成与全文件完成分开;旧结果没有坐标时不编造。工件只恢复已返回页,不能取得尚未返回的原源码。新发送视图和分页策略有版本,旧工件仍按其原视图核验。
分页按实际转义消息、下一请求固定开销、保留源码和同批实际用量分配。优化仅作用于纯文件读;混合操作和工件读仍采用原保守增长预留。紧张容量下匹配完整小源码保留边界,空间足够时继续支持更大页。1600 安全预留、完整目标/追加约束、工具配对、最终上下文和观察用量预留检查保持。
预算暂停分别展示窗口是否容纳与下一请求所需/余额。R 原 12352 可用输入、29623 预留和 17429 余额能直接从旧已存记录准确显示;27164 上游输入及 24192 缓存报告没有改写。显式续接仍只准备新 Run 的预览,不自动补预算或重放原工具。
## 验证中学到的事
- 更大的返回页不一定进入下一请求。首回放和相关测试发现重复定位开销、工件大页挤掉完整小源码;删掉重复定位,保留工件保守策略后改善。统一强制 2000 字符也会损害宽裕大文件读取,最终按可用容量采用边界。首失败日志保持。
- 同一历史证据不能按新版本重解释。工作依据与最新反馈的发送视图版本参与封存,旧版本断言继续核对;新测试从实际准备的视图取得版本,没有重写旧材料。
- 回放脚本也要核对真实契约。首计划 ID 选取、混合批次整页可见判据和 UI 用量位置有误,修正后保留原失败。后续取证助手误用 R 清单字段亦先拒绝,再读实际 files/path。旧实现 225 输入已逐字节匹配原 Git,R 原目标、输入、前三响应和原测试字节再次绑定。
- 发行要验证实际程序。独立源码包 565 输入重建 22 份编译输出,另核对 560 跟踪输入和 5 个示例生成输入;新 UI 分别验证源码、便携和实际安装副本。安装故障、窗口/请求/审批保护与正式保留原宿主分别验收。
- 首便携启动被 HOST_PORT_UNVERIFIED 拦下。确认原隔离目录为空且无所属宿主后,新隔离样本通过;保留首失败,不声称已证明端口冲突或某个系统原因,不修改保护以通过验收。
- 原生窗口最小化和用户输入变化导致操作拒绝;用户恢复后,按新状态观察再关闭,原宿主保持。原生导出实际为 JPEG,封存前仅纠正扩展名,未转换原图。
- 正式回看脚本的旧输入框选择器、窄屏隐藏侧栏可见等待和弹窗加载竞争分别暴露为验收器失败。保留每次结果,改为当前可访问名称、连接状态存在和实际池名称加载等待后,两视口通过。正式原生入口首次返回尚无可操作窗口,随后只读清单找到唯一实际窗口,没有重复启动。全程未为这些脚本问题改产品或重装。
## 结论和下一门槛
相同原目标/16K/2048/8 请求/90000 tokens、R 前三响应及固定续读策略下,旧实现 8 请求仍只在下一请求显示到 2173/2584,新实现 6 请求读完 2584 且纯文件页完整可见。原九测试仍失败、build 通过,0 源码修改和 Git 交付;这证明确定性读取流程改善,未证明真实模型自主交付或 Purge 收益。
正式交付身份、状态保留和原始结果以 [S 验收报告](QA-READING-RELIABILITY-S.md) 为准。下一按 [T 新样本计划](LIVE-DELIVERY-T-PLAN.md) 另冻结无工具身份预检和单个 Node 工程,保持原预算和监督边界;Node 通过后再 React/Python 及独立 Purge AB/BA。1.25 和重复完整目录核验性能独立排期,不为加速移除完整性与回退保护。
+2
View File
@@ -1,5 +1,7 @@
# Agent Workbench 验收记录
S已本地交付1.24.7/source5dcd6d7c/候选20261103164618190:文件续读携带原范围/偏移/SHA、纯文件批次容量及未用份额分配、Run/Task预留余额准确解释。96相关和18安装升级、源码/便携/实际安装新UI各10组、20组历史UI、565输入独立重建22/22、三归档/便携及12实际安装和16+9文件流程通过。正式原PID36552/Provider、63项目79会话6262记录13279事件及outbox/3660工件,所有原表行/全导出/工件与公开模型中心在网页/原生回看后再次精确保持,9229关闭。760份/106829868字节证据离线通过,见docs/QA-READING-RELIABILITY-S.md及docs/reading-s-release-identity.json。原R三响应+固定续读回放由8轮未完到6轮读完2584;原九测试仍失败,0修改/交付,本轮真实模型请求0。下一按docs/LIVE-DELIVERY-T-PLAN.md另冻结单次预检/新Node,16K/2048/8请求/90000tokens;不续跑旧失败,Node通过再React/Python/Purge。更新869425ms重复核验及1.25独立排期。不委派、不提取/复制/重存Key、不重启原Provider、不推送。
R真实实测与回顾已封存:唯一所属DNS/auto/IPv4/IPv6 TLS诊断通过,0HTTP/模型/凭据;新预检1请求/4492报告tokens/0工具/strict matched。新Node6请求/72571报告tokens/0未知、全JSON matched,12工具/1原检查审批,原九测试全部失败/基线build通过,0提案/修改/交付;第7次预留29623>剩余17429在发送前token_limit暂停,本地输入12352仍符合原上下文预算。原61项目77导出/3610工件/209尝试与原行(原项目位置按2个自有前插逆算)保持,预算/高性能池恢复,目标5→6→7/凭据1、全局65→69,9229关闭,当前63项目79会话6262记录13279事件及outbox/3660工件。178份/75944723字节证据及含清单179份/75980516字节Git索引核对通过;[本轮结果](QA-LIVE-DELIVERY-R.md)、[回顾](RETROSPECTIVE-LIVE-DELIVERY-R.md)与[下一S计划](READING-RELIABILITY-S-PLAN.md)已记录。正式仍1.24.6/source54dbe3d6/原PID36552/Provider,本轮没有产品改动、新发行或UI回看结论,React/Python/Purge未测。
Q既有发行记录:1.24.6/source54dbe3d6/候选20261103151145294已本地交付所属连接诊断。69相关/11宿主/14安装旧版升级与4补强热升级通过,560源码输入独立重建22/22、新UI源码/便携/安装各9组及五类历史26组、三归档/便携/12实际安装及16+9文件流程通过,原状态与正式网页/原生回看后再次保持。438份证据离线通过,真实模型0调用,完整验收见[Q报告](QA-NETWORK-DIAGNOSTICS-Q.md)及[发行身份](network-q-release-identity.json)。这是Q的历史验收,本轮R未重复执行这些检查;Q更新677755ms的重复核验性能继续单列。
@@ -0,0 +1,30 @@
{
"_approval-presentation-QyCkfhWb.js": {
"file": "assets/approval-presentation-QyCkfhWb.js",
"name": "approval-presentation"
},
"index.html": {
"file": "assets/desktop-CwtjTp8K.js",
"name": "desktop",
"src": "index.html",
"isEntry": true,
"imports": [
"_approval-presentation-QyCkfhWb.js"
],
"css": [
"assets/desktop-DC8YSCxQ.css"
]
},
"mobile.html": {
"file": "assets/mobile-DdtBN08G.js",
"name": "mobile",
"src": "mobile.html",
"isEntry": true,
"imports": [
"_approval-presentation-QyCkfhWb.js"
],
"css": [
"assets/mobile-joLHctQw.css"
]
}
}
File diff suppressed because one or more lines are too long
File diff suppressed because one or more lines are too long
File diff suppressed because one or more lines are too long
File diff suppressed because one or more lines are too long
File diff suppressed because one or more lines are too long
@@ -0,0 +1 @@
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 64 64"><rect x="2" y="2" width="60" height="60" rx="14" fill="#154b4a"/><path d="M12 44l8-25h6l8 25h-6l-2-6H18l-2 6zm8-11h5l-2.5-8zm14-14h5l3 17 3-13h5l3 13 3-17h5l-5 25h-6l-2.5-13L45 44h-6z" fill="#fcfdfc"/></svg>

After

Width:  |  Height:  |  Size: 268 B

@@ -0,0 +1,8 @@
<!doctype html>
<html lang="zh-CN">
<head><meta charset="UTF-8" /><meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="theme-color" content="#154b4a" /><link rel="icon" type="image/svg+xml" href="/favicon.svg" /><title>Agent Workbench · 本地编程工作台</title> <script type="module" crossorigin src="/assets/desktop-CwtjTp8K.js"></script>
<link rel="modulepreload" crossorigin href="/assets/approval-presentation-QyCkfhWb.js">
<link rel="stylesheet" crossorigin href="/assets/desktop-DC8YSCxQ.css">
</head>
<body><div id="root"></div></body>
</html>
@@ -0,0 +1,5 @@
[
"/assets/approval-presentation-QyCkfhWb.js",
"/assets/mobile-DdtBN08G.js",
"/assets/mobile-joLHctQw.css"
]
@@ -0,0 +1 @@
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 192 192"><rect width="192" height="192" rx="36" fill="#172b43"/><path d="M42 55h108v82H42z" fill="none" stroke="#f3f6fa" stroke-width="9"/><path d="m60 78 16 18-16 18m32 0h35" fill="none" stroke="#8bd0b9" stroke-width="9" stroke-linecap="square"/></svg>

After

Width:  |  Height:  |  Size: 307 B

@@ -0,0 +1,12 @@
const CACHE = 'workbench-mobile-65469da75610f50cf284';
const FILES = ["/","/mobile.webmanifest","/mobile-icon.svg","/mobile-touch-icon.png","/assets/approval-presentation-QyCkfhWb.js","/assets/mobile-DdtBN08G.js","/assets/mobile-joLHctQw.css"];
self.addEventListener('install', event => event.waitUntil(caches.open(CACHE).then(cache => cache.addAll(FILES)).then(() => self.skipWaiting())));
self.addEventListener('activate', event => event.waitUntil(caches.keys().then(names => Promise.all(names.filter(name => name.startsWith('workbench-mobile-') && name !== CACHE).map(name => caches.delete(name)))).then(() => self.clients.claim())));
self.addEventListener('fetch', event => {
const url = new URL(event.request.url);
if (event.request.method !== 'GET' || url.origin !== self.location.origin || url.search || !FILES.includes(url.pathname)) return;
event.respondWith(fetch(event.request).then(response => {
if (response.ok) { const copy = response.clone(); event.waitUntil(caches.open(CACHE).then(cache => cache.put(event.request, copy))); }
return response;
}).catch(async () => (await caches.match(event.request)) || Response.error()));
});
Binary file not shown.

After

Width:  |  Height:  |  Size: 1.0 KiB

@@ -0,0 +1,5 @@
<!doctype html>
<html lang="zh-CN"><head><meta charset="UTF-8" /><meta name="viewport" content="width=device-width, initial-scale=1, viewport-fit=cover" /><meta name="theme-color" content="#172b43" /><meta name="apple-mobile-web-app-capable" content="yes" /><meta name="apple-mobile-web-app-title" content="Workbench" /><link rel="manifest" href="/mobile.webmanifest" /><link rel="icon" type="image/svg+xml" href="/mobile-icon.svg" /><link rel="apple-touch-icon" href="/mobile-touch-icon.png" /><title>Agent Workbench · 手机伴随端</title> <script type="module" crossorigin src="/assets/mobile-DdtBN08G.js"></script>
<link rel="modulepreload" crossorigin href="/assets/approval-presentation-QyCkfhWb.js">
<link rel="stylesheet" crossorigin href="/assets/mobile-joLHctQw.css">
</head><body><div id="mobile-root"></div></body></html>
@@ -0,0 +1 @@
{"id":"/","name":"Agent Workbench 手机伴随端","short_name":"Workbench","lang":"zh-CN","start_url":"/","scope":"/","display":"standalone","background_color":"#f1f4f7","theme_color":"#172b43","icons":[{"src":"/mobile-icon.svg","sizes":"any","type":"image/svg+xml","purpose":"any"},{"src":"/mobile-touch-icon.png","sizes":"256x256","type":"image/png","purpose":"any"}]}
Binary file not shown.

After

Width:  |  Height:  |  Size: 1.1 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.0 KiB

File diff suppressed because it is too large Load Diff
@@ -0,0 +1,15 @@
"use strict";
// src/desktop/preload.cjs
var { contextBridge, ipcRenderer } = require("electron");
contextBridge.exposeInMainWorld("desktopWorkbench", Object.freeze({
protocol: "agent-workbench-desktop-v1",
status: () => ipcRenderer.invoke("awb:desktop-status"),
pickProject: () => ipcRenderer.invoke("awb:desktop-pick-project"),
projectFolder: (projectId) => ipcRenderer.invoke("awb:desktop-project-folder", { projectId }),
openFilePreview: (input) => ipcRenderer.invoke("awb:desktop-file-preview", input),
reviewTask: (input) => ipcRenderer.invoke("awb:desktop-review-task", input),
revealFile: (input) => ipcRenderer.invoke("awb:desktop-reveal-file", input),
stopHost: () => ipcRenderer.invoke("awb:desktop-stop-host"),
browsingLayout: (input) => ipcRenderer.invoke("awb:browsing-layout", input)
}));
@@ -0,0 +1,453 @@
// src/desktop/host-cli.mjs
import fs5 from "node:fs";
import path6 from "node:path";
import os from "node:os";
import { fileURLToPath } from "node:url";
// src/core/desktop/host-controller.mjs
import fs4 from "node:fs";
import path5 from "node:path";
import { spawn, execFileSync as execFileSync2 } from "node:child_process";
import { randomBytes, randomUUID as randomUUID2 } from "node:crypto";
import { setTimeout as delay } from "node:timers/promises";
// src/shared/contracts/desktop-host.mjs
import { createHmac, timingSafeEqual } from "node:crypto";
var HOST_PROTOCOL = "agent-workbench-host-v1";
var HOST_REGISTRATION_FILE = "host-authority-v1.json";
var hostError = (message, code = "HOST_IDENTITY_REJECTED", status = 409) => Object.assign(new Error(message), { code, status });
var canonicalHost = (value) => Array.isArray(value) ? value.map(canonicalHost) : value && typeof value === "object" ? Object.fromEntries(Object.keys(value).sort().map((key) => [key, canonicalHost(value[key])])) : value;
function hostProof(identity, challenge, secret) {
if (!/^[a-f0-9]{64}$/.test(challenge) || !Buffer.isBuffer(secret) || secret.length !== 32) throw hostError("\u5BBF\u4E3B\u6838\u9A8C\u6311\u6218\u683C\u5F0F\u65E0\u6548\u3002", "HOST_CHALLENGE_INVALID", 400);
return createHmac("sha256", secret).update(JSON.stringify(canonicalHost({ identity, challenge }))).digest("hex");
}
function verifyHostProof(reply, challenge, secret) {
if (!reply?.identity || reply.challenge !== challenge || !/^[a-f0-9]{64}$/.test(reply.proof ?? "")) throw hostError("\u5BBF\u4E3B\u6CA1\u6709\u63D0\u4F9B\u5339\u914D\u7684\u5B9E\u4F8B\u6838\u9A8C\u3002");
const actual = Buffer.from(reply.proof, "hex"), expected = Buffer.from(hostProof(reply.identity, challenge, secret), "hex");
if (!timingSafeEqual(actual, expected)) throw hostError("\u7AEF\u53E3\u54CD\u5E94\u4E0D\u5C5E\u4E8E\u767B\u8BB0\u7684\u6267\u884C\u5BBF\u4E3B\u3002");
return reply.identity;
}
// src/core/process-environment.mjs
import path from "node:path";
function projectEnvironment(extra = {}) {
const allowed = /^(path|pathext|systemroot|systemdrive|windir|comspec|temp|tmp|userprofile|homedrive|homepath|home|appdata|localappdata|programfiles(?:\(x86\))?|commonprogramfiles(?:\(x86\))?|lang|lc_all)$/i;
const environment = Object.fromEntries(Object.entries(process.env).filter(([name]) => allowed.test(name)));
const inheritedPath = Object.entries(environment).find(([name]) => name.toLowerCase() === "path")?.[1] || "";
for (const name of Object.keys(environment)) if (name.toLowerCase() === "path") delete environment[name];
return { ...environment, PATH: [path.dirname(process.execPath), inheritedPath].filter(Boolean).join(path.delimiter), ...extra };
}
// src/core/desktop/process-identity.mjs
import fs2 from "node:fs";
import path3 from "node:path";
import { execFileSync } from "node:child_process";
import { createHash } from "node:crypto";
// src/core/desktop/file-state.mjs
import fs from "node:fs";
import path2 from "node:path";
import { randomUUID } from "node:crypto";
var normalized = (value) => process.platform === "win32" ? value.toLowerCase() : value;
var samePath = (a, b) => normalized(path2.resolve(a)) === normalized(path2.resolve(b));
function assertLocalTree(absolute, { missing = false } = {}) {
const resolved = path2.resolve(absolute), parsed = path2.parse(resolved);
let current = parsed.root;
const components = resolved.slice(parsed.root.length).split(path2.sep).filter(Boolean);
for (const component of components) {
if (component.includes(":") || component.includes("\0")) throw hostError("\u672C\u673A\u72B6\u6001\u8DEF\u5F84\u683C\u5F0F\u65E0\u6548\u3002");
current = path2.join(current, component);
let stat;
try {
stat = fs.lstatSync(current);
} catch (cause) {
if (missing && cause.code === "ENOENT") return resolved;
throw cause;
}
if (stat.isSymbolicLink() || !samePath(current, fs.realpathSync.native(current))) throw hostError("\u672C\u673A\u72B6\u6001\u4E0D\u63A5\u53D7\u94FE\u63A5\u6216\u91CD\u89E3\u6790\u76EE\u5F55\u3002");
if (stat.isFile() && stat.nlink !== 1) throw hostError("\u672C\u673A\u72B6\u6001\u6587\u4EF6\u5177\u6709\u591A\u4E2A\u8DEF\u5F84\u3002");
}
return resolved;
}
function ownedChild(root, relative, { missing = false } = {}) {
if (typeof relative !== "string") throw hostError("\u672C\u673A\u72B6\u6001\u8DEF\u5F84\u65E0\u6548\u3002");
assertLocalTree(root);
const base = fs.realpathSync.native(root), target = path2.resolve(base, relative);
if (path2.isAbsolute(relative) || path2.win32.isAbsolute(relative) || relative.includes("\0") || !normalized(target).startsWith(normalized(base + path2.sep))) throw hostError("\u672C\u673A\u72B6\u6001\u8DEF\u5F84\u8D85\u51FA\u6240\u5C5E\u76EE\u5F55\u3002");
let existing = target;
while (!fs.existsSync(existing)) {
if (!missing) throw hostError("\u672C\u673A\u72B6\u6001\u6587\u4EF6\u4E0D\u5B58\u5728\u3002", "HOST_REGISTRATION_MISSING");
const parent = path2.dirname(existing);
if (parent === existing) throw hostError("\u672C\u673A\u72B6\u6001\u8DEF\u5F84\u65E0\u6CD5\u6838\u5BF9\u3002");
existing = parent;
}
const real = fs.realpathSync.native(existing);
if (!samePath(base, real) && !normalized(real).startsWith(normalized(base + path2.sep))) throw hostError("\u672C\u673A\u72B6\u6001\u94FE\u63A5\u8D85\u51FA\u6240\u5C5E\u76EE\u5F55\u3002");
if (fs.lstatSync(existing).isSymbolicLink()) throw hostError("\u672C\u673A\u72B6\u6001\u4E0D\u63A5\u53D7\u94FE\u63A5\u6216\u91CD\u89E3\u6790\u76EE\u5F55\u3002");
const stat = fs.statSync(existing);
if (stat.isFile() && stat.nlink !== 1) throw hostError("\u672C\u673A\u72B6\u6001\u6587\u4EF6\u5177\u6709\u591A\u4E2A\u8DEF\u5F84\u3002");
return assertLocalTree(target, { missing });
}
function readState(file, maximum = 4 * 1024 * 1024) {
assertLocalTree(file);
const stat = fs.lstatSync(file);
if (!stat.isFile() || stat.isSymbolicLink() || stat.nlink !== 1 || stat.size > maximum) throw hostError("\u672C\u673A\u72B6\u6001\u6587\u4EF6\u683C\u5F0F\u6216\u4F5C\u7528\u8303\u56F4\u65E0\u6548\u3002");
try {
return JSON.parse(fs.readFileSync(file, "utf8").replace(/^\uFEFF/, ""));
} catch {
throw hostError("\u672C\u673A\u72B6\u6001\u6587\u4EF6\u635F\u574F\uFF0C\u8BF7\u6838\u5BF9\u539F\u8BB0\u5F55\u3002");
}
}
function writeState(file, value) {
assertLocalTree(file, { missing: true });
const directory = path2.dirname(file), temporary = path2.join(directory, "." + path2.basename(file) + "-" + randomUUID() + ".tmp");
if (fs.existsSync(file)) readState(file);
try {
const descriptor = fs.openSync(temporary, "wx", 384);
try {
fs.writeFileSync(descriptor, JSON.stringify(value, null, 2) + "\n");
fs.fsyncSync(descriptor);
} finally {
fs.closeSync(descriptor);
}
assertLocalTree(file, { missing: true });
fs.renameSync(temporary, file);
} finally {
if (fs.existsSync(temporary)) fs.unlinkSync(temporary);
}
}
// src/core/desktop/process-identity.mjs
function windowsProcessObservationScript(pid) {
if (!Number.isInteger(pid) || pid < 1 || pid > 2147483647) throw hostError("\u5BBF\u4E3BPID\u65E0\u6548\u3002");
const query = 'Get-CimInstance Win32_Process -Filter "ProcessId=' + pid + '"';
return "$ErrorActionPreference='Stop';$ProgressPreference='SilentlyContinue';[Console]::OutputEncoding=New-Object System.Text.UTF8Encoding($false);$awbProcess=" + query + ';if($null -eq $awbProcess){Write-Output "null";exit 0};try{$awbOwner=Invoke-CimMethod -InputObject $awbProcess -MethodName GetOwnerSid;if($awbOwner.ReturnValue -ne 0 -or -not $awbOwner.Sid){throw "Owner identity unavailable"};}catch{$awbRemaining=' + query + ';if($null -eq $awbRemaining){Write-Output "null";exit 0};throw};[pscustomobject]@{pid=[int]$awbProcess.ProcessId;parentPid=[int]$awbProcess.ParentProcessId;createdAt=$awbProcess.CreationDate.ToUniversalTime().ToString("yyyy-MM-ddTHH:mm:ss.fffZ");executable=$awbProcess.ExecutablePath;userSid=$awbOwner.Sid}|ConvertTo-Json -Compress';
}
function observeWindowsProcess(pid) {
if (process.platform !== "win32") throw hostError("\u6B64\u684C\u9762\u5BBF\u4E3B\u6838\u9A8C\u9700\u8981Windows\u3002", "HOST_PLATFORM_UNSUPPORTED");
const script = windowsProcessObservationScript(pid);
try {
const shell = path3.join(process.env.SystemRoot || "C:\\Windows", "System32/WindowsPowerShell/v1.0/powershell.exe");
const output = execFileSync(
shell,
["-NoProfile", "-NonInteractive", "-EncodedCommand", Buffer.from(script, "utf16le").toString("base64")],
{ windowsHide: true, encoding: "utf8", stdio: ["pipe", "pipe", "pipe"], timeout: 15e3, maxBuffer: 65536, env: projectEnvironment() }
);
const value = JSON.parse(output.trim().replace(/^\uFEFF/, ""));
if (value === null) return null;
if (value.pid !== pid || !Number.isFinite(Date.parse(value.createdAt)) || !/^S-1-\d+(?:-\d+)+$/.test(value.userSid) || !value.executable || !fs2.existsSync(value.executable)) throw new Error("Invalid observation");
return { ...value, executable: fs2.realpathSync.native(value.executable) };
} catch {
throw hostError("\u65E0\u6CD5\u6838\u5BF9Windows\u5BBF\u4E3B\u521B\u5EFA\u3001\u7528\u6237\u6216\u53EF\u6267\u884C\u8EAB\u4EFD\uFF1B\u4E0D\u4F1A\u590D\u7528\u6216\u505C\u6B62\u672A\u77E5\u8FDB\u7A0B\u3002", "HOST_PROCESS_UNVERIFIED");
}
}
function binaryIdentity(file) {
assertLocalTree(file);
const absolute = fs2.realpathSync.native(file), stat = fs2.lstatSync(absolute);
if (!stat.isFile() || stat.isSymbolicLink() || stat.nlink !== 1 || stat.size > 256 * 1024 * 1024) throw hostError("\u5BBF\u4E3B\u53EF\u6267\u884C\u6587\u4EF6\u8EAB\u4EFD\u65E0\u6548\u3002");
return { path: absolute, bytes: stat.size, sha256: createHash("sha256").update(fs2.readFileSync(absolute)).digest("hex") };
}
// src/core/desktop/app-installation.mjs
import fs3 from "node:fs";
import path4 from "node:path";
import { createHash as createHash2 } from "node:crypto";
function inspectHostInstallation(directory, { full = false, onProgress = null } = {}) {
assertLocalTree(directory);
const root = fs3.realpathSync.native(path4.resolve(directory));
if (!fs3.statSync(root).isDirectory()) throw hostError("\u5BBF\u4E3B\u5B89\u88C5\u76EE\u5F55\u65E0\u6548\u3002");
const digest = (value) => createHash2("sha256").update(value).digest("hex");
const manifestFile = path4.join(root, "MANIFEST.json"), manifest = fs3.existsSync(manifestFile) ? readState(manifestFile, 16 * 1024 * 1024) : null;
if (manifest && (manifest.app !== "agent-workbench" || !Array.isArray(manifest.files) || manifest.files.length > 6e4)) throw hostError("\u5BBF\u4E3B\u53D1\u884C\u6E05\u5355\u65E0\u6548\u3002");
const version = manifest?.version ?? (fs3.existsSync(path4.join(root, "package.json")) ? readState(path4.join(root, "package.json")).version : null);
if (typeof version !== "string" || !/^\d+\.\d+\.\d+$/.test(version)) throw hostError("\u5BBF\u4E3B\u5B89\u88C5\u7248\u672C\u65E0\u6548\u3002");
const files = /* @__PURE__ */ new Set(["dist-server/server.mjs", "dist-server/hot-upgrade.mjs", "dist-client/index.html"]);
if (!manifest || fs3.existsSync(path4.join(root, "dist-server/host-launcher.mjs"))) files.add("dist-server/host-launcher.mjs");
if (fs3.existsSync(path4.join(root, "dist-server/install-manager.mjs"))) files.add("dist-server/install-manager.mjs");
for (const relative of ["dist-desktop/main.cjs", "dist-desktop/preload.cjs"]) if (fs3.existsSync(path4.join(root, relative))) files.add(relative);
const names = /* @__PURE__ */ new Map();
for (const row of manifest?.files ?? []) {
if (!row || typeof row.path !== "string" || row.path.includes("\\") || row.path.split("/").includes("..") || names.has(row.path) || !/^[a-f0-9]{64}$/.test(row.sha256 ?? "") || !Number.isSafeInteger(row.bytes) || row.bytes < 0) throw hostError("\u5BBF\u4E3B\u6E05\u5355\u5305\u542B\u91CD\u590D\u6216\u8D8A\u754C\u8F93\u5165\u3002");
names.set(row.path, row);
if (full || /^(dist-client|dist-server|dist-desktop)\//.test(row.path)) files.add(row.path);
}
if (!manifest || manifest.kind === "source") for (const entry of fs3.readdirSync(path4.join(root, "dist-client/assets"))) files.add("dist-client/assets/" + entry);
const selected = [...files].sort(), totalBytes = selected.reduce((sum, name) => sum + (names.get(name)?.bytes ?? 0), 0);
let completedFiles = 0, completedBytes = 0;
const report = (currentPath) => onProgress?.({ scope: full ? "full-manifest" : "loaded-inputs", root, totalFiles: selected.length, totalBytes: manifest ? totalBytes : null, completedFiles, completedBytes, currentPath });
report(null);
const inputs = selected.map((relative) => {
const file = ownedChild(root, relative), identity = binaryIdentity(file), declared = names.get(relative);
if (manifest?.kind === "portable" && !declared) throw hostError("\u5B9E\u9645\u5BBF\u4E3B\u7A0B\u5E8F\u4E0D\u5728\u539F\u53D1\u884C\u6E05\u5355\u4E2D\u3002");
if (declared && (identity.bytes !== declared.bytes || identity.sha256 !== declared.sha256)) throw hostError("\u5B9E\u9645\u5BBF\u4E3B\u7A0B\u5E8F\u4E0E\u539F\u53D1\u884C\u6E05\u5355\u4E0D\u4E00\u81F4\u3002");
completedFiles++;
completedBytes += identity.bytes;
report(relative);
return { path: relative, bytes: identity.bytes, sha256: identity.sha256 };
});
const result = {
root,
version,
kind: manifest?.kind ?? "development-build",
manifestSha256: manifest ? digest(fs3.readFileSync(manifestFile)) : null,
inputs,
verification: full ? "full-manifest" : "loaded-host-and-client-inputs"
};
return { ...result, identitySha256: digest(JSON.stringify(canonicalHost(result))) };
}
// src/core/desktop/host-controller.mjs
var HOST_ENVIRONMENT = Object.freeze(["AGENT_WORKBENCH_DEEPSEEK_ROOT", "AGENT_WORKBENCH_PURGE_ROOT", "AGENT_WORKBENCH_PYTHON_PATH", "AGENT_WORKBENCH_TEST_PYTHON"]);
async function jsonResponse(response) {
if (!response.ok) throw hostError("\u672C\u673A\u5BBF\u4E3B\u63A5\u53E3\u62D2\u7EDD\u64CD\u4F5C\uFF08HTTP " + response.status + "\uFF09\u3002", "HOST_API_REJECTED", response.status);
const chunks = [];
let length = 0;
for await (const chunk of response.body) {
length += chunk.length;
if (length > 4 * 1024 * 1024) throw hostError("\u672C\u673A\u5BBF\u4E3B\u54CD\u5E94\u8FC7\u5927\u3002");
chunks.push(chunk);
}
try {
return JSON.parse(Buffer.concat(chunks).toString("utf8"));
} catch {
throw hostError("\u7AEF\u53E3\u672A\u8FD4\u56DE\u6709\u6548\u7684\u5BBF\u4E3B\u534F\u8BAE\u3002");
}
}
var HostController = class {
constructor({ appRoot: appRoot2, dataRoot: dataRoot2, port = 4318, nodePath: nodePath2 = null, timeoutMs = 45e3, environment = {} }) {
if (!Number.isInteger(port) || port < 1 || port > 65535 || typeof dataRoot2 !== "string" || !path5.isAbsolute(dataRoot2)) throw hostError("\u684C\u9762\u5BBF\u4E3B\u76EE\u5F55\u6216\u7AEF\u53E3\u65E0\u6548\u3002");
this.installation = inspectHostInstallation(appRoot2);
this.appRoot = this.installation.root;
assertLocalTree(dataRoot2, { missing: true });
fs4.mkdirSync(dataRoot2, { recursive: true });
this.dataRoot = fs4.realpathSync.native(dataRoot2);
const bundled = path5.join(this.appRoot, "runtime/node.exe");
const executable = nodePath2 ?? (fs4.existsSync(bundled) ? bundled : !process.versions.electron ? process.execPath : null);
if (!executable) throw hostError("\u684C\u9762\u5B89\u88C5\u7F3A\u5C11\u72EC\u7ACBNode\u5BBF\u4E3B\u3002");
this.node = binaryIdentity(executable);
const version = execFileSync2(this.node.path, ["--version"], { encoding: "utf8", windowsHide: true, timeout: 1e4, env: projectEnvironment() }).trim();
if (!/^v24\./.test(version)) throw hostError("\u684C\u9762\u5BBF\u4E3B\u9700\u8981\u5DF2\u7ECF\u6838\u9A8C\u7684Node24\u8FD0\u884C\u5668\u3002");
this.node.version = version;
this.port = port;
this.url = "http://127.0.0.1:" + port;
this.timeoutMs = Math.min(6e4, Math.max(1e3, timeoutMs));
const allowed = new Set(HOST_ENVIRONMENT);
if (Object.keys(environment).some((key) => !allowed.has(key) || typeof environment[key] !== "string" || !path5.isAbsolute(environment[key]))) throw hostError("\u5BBF\u4E3B\u542F\u52A8\u53EA\u63A5\u53D7\u5DF2\u58F0\u660E\u7684\u672C\u673A\u8FD0\u884C\u5668\u8DEF\u5F84\u3002");
this.environment = environment;
}
async api(route, { method = "GET", input = null, timeoutMs = 15e3 } = {}) {
if (typeof route !== "string" || !route.startsWith("/") || route.startsWith("//")) throw hostError("\u5BBF\u4E3B\u63A5\u53E3\u8DEF\u5F84\u65E0\u6548\u3002");
return jsonResponse(await fetch(this.url + "/api" + route, {
method,
headers: {
Connection: "close",
...this.token ? { "X-Workbench-Token": this.token } : {},
...input ? { "Content-Type": "application/json" } : {}
},
body: input ? JSON.stringify(input) : void 0,
signal: AbortSignal.timeout(timeoutMs)
}));
}
async health() {
try {
return await this.api("/health", { timeoutMs: 1500 });
} catch (cause) {
if (cause.cause?.code === "ECONNREFUSED" || cause.cause?.errors?.every((value) => value.code === "ECONNREFUSED")) return null;
throw hostError("\u7AEF\u53E3\u88AB\u5360\u7528\u6216\u54CD\u5E94\u65E0\u6CD5\u6838\u5BF9\uFF1B\u4E0D\u4F1A\u542F\u52A8\u53E6\u4E00\u5957\u5BBF\u4E3B\u3002", "HOST_PORT_UNVERIFIED");
}
}
async connect() {
const health = await this.health();
if (!health) return null;
if (health.app !== "agent-workbench" || health.ready !== true || !Number.isInteger(health.pid)) throw hostError("\u7AEF\u53E3\u5C5E\u4E8E\u672A\u77E5\u670D\u52A1\u3002");
const bootstrap = await this.api("/bootstrap");
if (bootstrap.version !== this.installation.version || typeof bootstrap.token !== "string") throw hostError("\u8FD0\u884C\u7248\u672C\u4E0D\u5339\u914D\uFF0C\u8BF7\u4F7F\u7528\u660E\u786E\u7684\u5347\u7EA7\u6D41\u7A0B\u3002", "HOST_VERSION_MISMATCH");
this.token = bootstrap.token;
const challenge = randomBytes(32).toString("hex"), reply = await this.api("/host/identity?challenge=" + challenge);
ownedChild(this.dataRoot, ".secrets");
const registration = readState(ownedChild(this.dataRoot, ".secrets/" + HOST_REGISTRATION_FILE), 16384);
if (registration.schema !== 1 || registration.protocol !== HOST_PROTOCOL || !/^[A-Za-z0-9+/]{43}=$/.test(registration.secret ?? "")) throw hostError("\u672C\u673A\u539F\u6CE8\u518C\u683C\u5F0F\u65E0\u6CD5\u6838\u5BF9\u3002");
const identity = verifyHostProof(reply, challenge, Buffer.from(registration.secret, "base64"));
const actual = observeWindowsProcess(identity.pid), caller = observeWindowsProcess(process.pid);
if (!actual || !caller || identity.pid !== health.pid || identity.pid !== registration.pid || identity.createdAt !== actual.createdAt || identity.createdAt !== registration.createdAt || identity.instanceId !== registration.instanceId || identity.ownerId !== registration.ownerId || identity.userSid !== actual.userSid || identity.userSid !== registration.userSid || identity.userSid !== caller.userSid || identity.protocol !== HOST_PROTOCOL || identity.version !== this.installation.version || identity.port !== this.port || !samePath(identity.dataRoot, this.dataRoot) || !samePath(identity.runtimeRoot, this.appRoot) || identity.runtimeIdentity !== this.installation.identitySha256 || !samePath(identity.executable.path, actual.executable) || identity.executable.sha256 !== this.node.sha256 || binaryIdentity(actual.executable).sha256 !== this.node.sha256) throw hostError("\u5B9E\u9645PID\u521B\u5EFA\u3001\u7528\u6237\u3001\u7A0B\u5E8F\u3001\u5B89\u88C5\u6216\u6570\u636E\u76EE\u5F55\u4E0E\u539F\u5BBF\u4E3B\u767B\u8BB0\u4E0D\u4E00\u81F4\u3002");
const info = readState(ownedChild(this.dataRoot, "service-info.json"));
if (info.pid !== identity.pid || info.instanceId !== identity.instanceId || info.createdAt !== identity.createdAt || info.port !== this.port || info.userSid !== identity.userSid || info.ownerId !== identity.ownerId || info.runtimeIdentity !== identity.runtimeIdentity || !samePath(info.appRoot, identity.appRoot) || !samePath(info.dataRoot, this.dataRoot) || !samePath(info.runtimeRoot, this.appRoot)) throw hostError("\u672C\u673A\u5BBF\u4E3B\u7D22\u5F15\u4E0E\u5B9E\u9645\u5B9E\u4F8B\u4E0D\u4E00\u81F4\u3002");
this.identity = identity;
return { url: this.url, identity };
}
async lock(action) {
const lockDirectory = ownedChild(this.dataRoot, ".desktop-launch.lock", { missing: true }), nonce = randomUUID2(), observed = observeWindowsProcess(process.pid);
const deadline = Date.now() + this.timeoutMs;
while (true) {
try {
fs4.mkdirSync(lockDirectory, { mode: 448 });
break;
} catch (cause) {
if (cause.code !== "EEXIST") throw cause;
ownedChild(this.dataRoot, ".desktop-launch.lock");
const ownerFile2 = path5.join(lockDirectory, "owner.json");
if (fs4.existsSync(ownerFile2)) {
ownedChild(this.dataRoot, ".desktop-launch.lock/owner.json");
const owner = readState(ownerFile2, 16384);
if (!Number.isInteger(owner.pid) || typeof owner.createdAt !== "string" || typeof owner.nonce !== "string") throw hostError("\u542F\u52A8\u9501\u7684\u539F\u8EAB\u4EFD\u65E0\u6CD5\u6838\u5BF9\u3002");
const active = observeWindowsProcess(owner.pid);
if (active && active.createdAt === owner.createdAt && (owner.userSid !== active.userSid || owner.userSid !== observed.userSid)) throw hostError("\u542F\u52A8\u9501\u4E0D\u5C5E\u4E8E\u5F53\u524DWindows\u7528\u6237\u3002");
if (!active || active.createdAt !== owner.createdAt) {
if (readState(ownerFile2).nonce !== owner.nonce) continue;
const archive = ownedChild(this.dataRoot, ".desktop-launch-stale-" + randomUUID2(), { missing: true });
try {
fs4.renameSync(lockDirectory, archive);
} catch (error) {
if (!["ENOENT", "EEXIST"].includes(error.code)) throw error;
}
continue;
}
} else if (Date.now() - fs4.statSync(lockDirectory).mtimeMs > 1e4) throw hostError("\u542F\u52A8\u9501\u7F3A\u5C11\u539F\u8BB0\u5F55\uFF0C\u8BF7\u6838\u5BF9\u8BE5\u76EE\u5F55\u3002");
if (Date.now() >= deadline) throw hostError("\u6B63\u5728\u7B49\u5F85\u539F\u542F\u52A8\u64CD\u4F5C\uFF0C\u8BF7\u6838\u5BF9\u5F53\u524D\u5BBF\u4E3B\u3002", "HOST_START_PENDING");
await delay(100);
}
}
const ownerFile = path5.join(lockDirectory, "owner.json");
try {
writeState(ownerFile, { schema: 1, pid: process.pid, createdAt: observed.createdAt, userSid: observed.userSid, nonce });
return await action();
} finally {
if (fs4.existsSync(ownerFile) && readState(ownerFile).nonce === nonce) {
fs4.unlinkSync(ownerFile);
fs4.rmdirSync(lockDirectory);
}
}
}
async start() {
return this.lock(async () => {
const existing = await this.connect();
if (existing) return { ...existing, reused: true };
const intentFile = ownedChild(this.dataRoot, "desktop-launch-state.json", { missing: true });
if (fs4.existsSync(intentFile)) {
const intent = readState(intentFile);
if (intent.pid && !["exited", "failed"].includes(intent.state)) {
const actual = observeWindowsProcess(intent.pid);
if (actual && actual.createdAt === intent.createdAt) throw hostError("\u539F\u542F\u52A8\u8FDB\u7A0B\u4ECD\u6D3B\u52A8\u4F46\u6CA1\u6709\u53EF\u6838\u9A8C\u63A5\u53E3\uFF0C\u8BF7\u6838\u5BF9\u539F\u65E5\u5FD7\uFF1B\u4E0D\u4F1A\u518D\u542F\u52A8\u53E6\u4E00\u8FDB\u7A0B\u3002", "HOST_START_UNVERIFIED");
}
}
const logs = ownedChild(this.dataRoot, "logs", { missing: true });
if (!fs4.existsSync(logs)) fs4.mkdirSync(logs);
const id = randomUUID2(), stdoutPath = path5.join(logs, "desktop-host-" + id + "-stdout.log"), stderrPath = path5.join(logs, "desktop-host-" + id + "-stderr.log");
let output, errors, child;
try {
output = fs4.openSync(stdoutPath, "wx", 384);
errors = fs4.openSync(stderrPath, "wx", 384);
child = spawn(this.node.path, [ownedChild(this.appRoot, "dist-server/server.mjs")], {
cwd: this.appRoot,
windowsHide: true,
detached: true,
env: projectEnvironment({ ...this.environment, AGENT_WORKBENCH_APP_DIR: this.appRoot, AGENT_WORKBENCH_DATA_DIR: this.dataRoot, AGENT_WORKBENCH_PORT: String(this.port) }),
stdio: ["ignore", output, errors]
});
} finally {
if (output !== void 0) fs4.closeSync(output);
if (errors !== void 0) fs4.closeSync(errors);
}
let failure = null;
child.on("error", (cause) => {
failure = cause;
});
child.unref();
const observedChild = child.pid ? observeWindowsProcess(child.pid) : null;
writeState(intentFile, {
schema: 1,
state: observedChild ? "starting" : "failed",
id,
pid: child.pid ?? null,
createdAt: observedChild?.createdAt ?? null,
appRoot: this.appRoot,
dataRoot: this.dataRoot,
port: this.port,
stdoutPath,
stderrPath
});
const deadline = Date.now() + this.timeoutMs;
while (Date.now() < deadline) {
if (failure || child.exitCode !== null) throw hostError("\u6240\u5C5E\u5BBF\u4E3B\u542F\u52A8\u5931\u8D25\uFF0C\u8BF7\u67E5\u770B\u672C\u673A\u542F\u52A8\u8BB0\u5F55\u3002", "HOST_START_FAILED");
await delay(100);
let health;
try {
health = await this.health();
} catch (cause) {
if (cause.code !== "HOST_PORT_UNVERIFIED") throw cause;
const waiting = observeWindowsProcess(child.pid);
if (!waiting || !observedChild || waiting.createdAt !== observedChild.createdAt || waiting.userSid !== observedChild.userSid || !samePath(waiting.executable, this.node.path)) throw hostError("\u6240\u5C5E\u5BBF\u4E3B\u542F\u52A8\u8FDB\u7A0B\u5DF2\u7ED3\u675F\u6216\u8EAB\u4EFD\u53D8\u5316\uFF0C\u8BF7\u6838\u5BF9\u539F\u65E5\u5FD7\u3002", "HOST_START_FAILED");
continue;
}
if (!health) continue;
if (health.pid !== child.pid) throw hostError("\u542F\u52A8\u671F\u95F4\u7AEF\u53E3\u88AB\u5176\u4ED6\u8FDB\u7A0B\u5360\u7528\u3002");
const connected = await this.connect();
if (connected) {
writeState(intentFile, { ...readState(intentFile), state: "ready", instanceId: connected.identity.instanceId });
return { ...connected, reused: false, logs: { stdoutPath, stderrPath } };
}
}
throw hostError("\u6240\u5C5E\u5BBF\u4E3B\u5C1A\u672A\u5B8C\u6210\u6CE8\u518C\uFF1B\u4E0D\u4F1A\u91CD\u65B0\u542F\u52A8\u6216\u7ED3\u675F\u672A\u77E5\u8FDB\u7A0B\u3002", "HOST_START_UNVERIFIED");
});
}
async status() {
if (!await this.connect()) return { state: "not-running", url: this.url };
return this.api("/host/status");
}
async stop() {
return this.lock(async () => {
const connection = await this.connect();
if (!connection) return { state: "not-running", url: this.url };
const original = connection.identity;
await this.api("/host/stop", { method: "POST", input: {
instanceId: original.instanceId,
createdAt: original.createdAt,
runtimeIdentity: original.runtimeIdentity,
confirmation: "stop-owned-host"
}, timeoutMs: this.timeoutMs });
const deadline = Date.now() + this.timeoutMs;
while (Date.now() < deadline) {
await delay(200);
const observed = observeWindowsProcess(original.pid);
if (!observed) {
if (await this.health()) throw hostError("\u539F\u5BBF\u4E3B\u5DF2\u7ED3\u675F\uFF0C\u4F46\u7AEF\u53E3\u5B58\u5728\u5176\u4ED6\u670D\u52A1\u3002");
return { state: "stopped", identity: original, processExited: true };
}
if (observed.createdAt !== original.createdAt) throw hostError("\u539FPID\u5DF2\u88AB\u590D\u7528\uFF0C\u4E0D\u4F1A\u7ED3\u675F\u65B0\u7684\u8FDB\u7A0B\u3002");
}
throw hostError("\u6240\u5C5E\u5BBF\u4E3B\u9000\u51FA\u4ECD\u5F85\u6838\u5BF9\uFF0C\u8BF7\u4FDD\u7559\u539F\u5B9E\u4F8B\u8BB0\u5F55\u3002", "HOST_STOP_UNVERIFIED");
});
}
};
// src/desktop/host-cli.mjs
var args = process.argv.slice(2);
if (args.includes("--help")) {
console.log("node host-launcher.mjs <start|status|stop> [--app-root <absolute>] [--data-root <absolute>] [--node <absolute>] [--port <1..65535>]");
process.exit(0);
}
var command = args.shift();
if (!["start", "status", "stop"].includes(command)) throw new Error("Use --help for a supported desktop host operation.");
var options = {};
while (args.length) {
const flag = args.shift(), value = args.shift();
if (!["--app-root", "--data-root", "--node", "--port"].includes(flag) || !value || options[flag] !== void 0) throw new Error("Invalid or repeated host option.");
options[flag] = value;
}
var defaultRoot = path6.resolve(path6.dirname(fileURLToPath(import.meta.url)), "..");
var appRoot = options["--app-root"] || process.env.AGENT_WORKBENCH_APP_DIR || defaultRoot;
var dataRoot = options["--data-root"] || process.env.AGENT_WORKBENCH_DATA_DIR || path6.join(process.env.LOCALAPPDATA || path6.join(os.homedir(), ".local/share"), "AgentWorkbench");
var nodePath = options["--node"] || (fs5.existsSync(path6.join(appRoot, "runtime/node.exe")) ? path6.join(appRoot, "runtime/node.exe") : process.execPath);
try {
const controller = new HostController({
appRoot,
dataRoot,
nodePath,
port: Number(options["--port"] || process.env.AGENT_WORKBENCH_PORT || 4318),
environment: Object.fromEntries(HOST_ENVIRONMENT.filter((key) => process.env[key]).map((key) => [key, process.env[key]]))
});
console.log(JSON.stringify(await controller[command]()));
} catch (cause) {
console.error(JSON.stringify({ success: false, code: cause.code || "HOST_FAILED", message: cause.message }));
process.exitCode = 1;
}
File diff suppressed because one or more lines are too long
File diff suppressed because it is too large Load Diff
File diff suppressed because one or more lines are too long
@@ -0,0 +1,62 @@
import fs from 'node:fs';
import path from 'node:path';
import assert from 'node:assert/strict';
import { createHash } from 'node:crypto';
import { DatabaseSync } from 'node:sqlite';
import { snapshotPersistentData } from '../src/core/desktop/install-manager.mjs';
export const hash = bytes => createHash('sha256').update(bytes).digest('hex');
export const canonical = value => Array.isArray(value) ? value.map(canonical) : value && typeof value === 'object' ? Object.fromEntries(Object.keys(value).sort().map(key => [key, canonical(value[key])])) : value;
export const digest = value => hash(JSON.stringify(canonical(value)));
export function databaseSnapshot(dataRoot) {
const database = new DatabaseSync(path.join(dataRoot, 'workbench.sqlite'), { readOnly: true }); database.exec('BEGIN');
try {
assert.equal(database.prepare('PRAGMA quick_check').get().quick_check, 'ok');
const tables = {};
for (const { name, sql } of database.prepare("SELECT name,sql FROM sqlite_schema WHERE type='table' ORDER BY name").all()) {
assert.match(name, /^[a-z_]+$/);
const rows = database.prepare('SELECT * FROM "' + name + '" ORDER BY rowid').all().filter(row => !(name === 'metadata' && row.key === 'runtimeOwner'));
tables[name] = { count: rows.length, schemaSha256: hash(sql), sha256: digest(rows), rows: rows.map((row, index) => ({ id: row.id ?? row.key ?? row.seq ?? index, ...(row.kind ? { kind: row.kind } : {}), sha256: digest(row) })) };
}
const sessionIds = database.prepare('SELECT id FROM sessions ORDER BY id').all().map(row => row.id);
const owner = database.prepare("SELECT value FROM metadata WHERE key='runtimeOwner'").get();
return { tables, sessionIds, ownerSha256: owner ? digest(owner) : null, scope: 'Every original table schema and row field hashed, excluding only runtimeOwner lifecycle metadata. No row payload or credential is returned.' };
} finally { database.exec('ROLLBACK'); database.close(); }
}
export async function publicSnapshot(url, dataRoot) {
const healthResponse = await fetch(url + '/api/health', { headers: { Connection: 'close' }, signal: AbortSignal.timeout(20000) }); assert.equal(healthResponse.status, 200);
const health = await healthResponse.json(), bootResponse = await fetch(url + '/api/bootstrap', { headers: { Connection: 'close' }, signal: AbortSignal.timeout(20000) }); assert.equal(bootResponse.status, 200);
const bootstrap = await bootResponse.json(); const token = bootstrap.token;
const api = async route => { const response = await fetch(url + '/api' + route, { headers: { 'X-Workbench-Token': token, Connection: 'close' }, signal: AbortSignal.timeout(20000) }); assert.ok(response.ok, route + ': HTTP ' + response.status); return response.json(); };
const database = databaseSnapshot(dataRoot), exports = [];
for (const id of database.sessionIds) {
const report = await api('/sessions/' + id + '/export');
exports.push({ id, project: digest(report.project), session: digest(report.session), execution: digest(report.execution) });
}
const center = await api('/model-center');
// This is the documented public model-center view; Provider's private map and
// API key are never accessed. Only this public object is retained for schema-aware comparison.
const modelCenter = structuredClone(center);
return { observedAt: new Date().toISOString(), version: health.version, pid: health.pid, runtimeRoot: bootstrap.runtimeRoot,
configSha256: digest(bootstrap.config), projectsSha256: digest(bootstrap.projects), sessionsSha256: digest(bootstrap.sessions), modelCenterSha256: digest(modelCenter), modelCenter,
projects: bootstrap.projects.length, visibleSessions: bootstrap.sessions.length, hasKey: !!bootstrap.config.hasKey,
active: bootstrap.sessions.filter(value => ['running', 'awaiting_approval', 'stopping'].includes(value.status)).map(value => ({ id: value.id, status: value.status })),
storage: bootstrap.storage, eventCursor: bootstrap.eventCursor, database, exports, persistent: snapshotPersistentData(dataRoot),
scope: 'Read-only public state, every original session export digest, original table/row/schema hashes and every original artifact byte hash. No credential extraction, model probe, execution or mutation.' };
}
export function compareOriginalState(before, after, { modelCenterSchemaChange = false } = {}) {
for (const field of ['pid', 'hasKey', 'configSha256', 'projectsSha256', 'sessionsSha256', 'eventCursor']) assert.deepEqual(after[field], before[field], field);
assert.deepEqual(canonical(after.database.tables), canonical(before.database.tables), 'All original table schemas and every persistent row');
assert.deepEqual(after.database.sessionIds, before.database.sessionIds); assert.deepEqual(after.exports, before.exports);
assert.deepEqual(after.persistent, before.persistent, 'Every original artifact byte, persistent row and public disk config');
if (modelCenterSchemaChange) {
const original = structuredClone(before.modelCenter), next = structuredClone(after.modelCenter);
// 1.15 changes only the public explanatory text about optional DPAPI storage.
// Existing IDs, routes, revisions, credentials-present bits and all original attempts remain exact.
assert.match(next.storage.restart, /临时凭据.*已明确本机加密保存/);
next.storage.restart = original.storage.restart;
assert.deepEqual(next, original, 'Original public model center, except versioned storage explanation');
return { modelCenterChange: { path: '/storage/restart', original: before.modelCenter.storage.restart, actual: after.modelCenter.storage.restart }, originalRows: 'exact', originalExports: 'exact', originalArtifactBytes: 'exact' };
}
assert.equal(after.modelCenterSha256, before.modelCenterSha256);
return { originalRows: 'exact', originalExports: 'exact', originalArtifactBytes: 'exact', originalModelCenter: 'exact' };
}
@@ -0,0 +1,16 @@
import fs from 'node:fs';
import assert from 'node:assert/strict';
let body = fs.readFileSync('.qa/network-q-formal-release.mjs', 'utf8');
const changes = [
['.qa/network-q/formal-release', '.qa/reading-reliability-s/formal-release'],
['.qa/network-q/initial-private-state.json', '.qa/reading-reliability-s/production-before.json'],
['1.24.6', '__NEXT_VERSION__'], ['1.24.5', '1.24.6'], ['__NEXT_VERSION__', '1.24.7'],
['1\\.24\\.5-win-x64-20261103133956264', '1\\.24\\.6-win-x64-20261103151145294']
];
for (const [before, after] of changes) { assert.ok(body.includes(before), before); body = body.replaceAll(before, after); }
const line = "const archived = read('.qa/archives-validation.json'), portable = read('.qa/release-validation.json');";
assert.ok(body.includes(line));
body = body.replace(line,
"assert.equal(read('.qa/reading-reliability-s/release/preinstall-review.json').success, true);\n " + line);
fs.writeFileSync('.qa/reading-s-formal-release.mjs', body, { flag: 'wx' });
console.log('Prepared S retained-host formal installer driver.');
@@ -0,0 +1,19 @@
import fs from 'node:fs';
import path from 'node:path';
import assert from 'node:assert/strict';
import { publicSnapshot } from './desktop-f-state.mjs';
const root = path.resolve('.qa/reading-reliability-s');
fs.mkdirSync(root, { recursive: true });
const destination = path.join(root, 'production-before.json');
assert.ok(!fs.existsSync(destination), 'S baseline is immutable');
const data = 'C:/Users/wangy/AppData/Local/Packages/OpenAI.Codex_2p2nqsd0c76g0/LocalCache/Local/AgentWorkbench';
const state = await publicSnapshot('http://127.0.0.1:4318', data);
assert.equal(state.pid, 36552);
assert.equal(state.version, '1.24.6');
assert.deepEqual(state.active, []);
fs.writeFileSync(destination, JSON.stringify(state, null, 2) + '\n');
console.log(JSON.stringify({ version: state.version, pid: state.pid, hasKey: state.hasKey,
projects: state.projects, sessions: state.database.sessionIds.length,
rows: Object.fromEntries(Object.entries(state.database.tables).map(([key, value]) => [key, value.count])),
artifactFiles: state.persistent.artifacts?.length, active: state.active }));
@@ -0,0 +1,50 @@
import fs from 'node:fs';
import path from 'node:path';
import assert from 'node:assert/strict';
import { chromium } from 'playwright';
const root = path.resolve('.qa/reading-reliability-s/release/browser-review');
assert.equal(fs.existsSync(root), false); fs.mkdirSync(root);
const origin = 'http://127.0.0.1:4318', errors = [], rejectedWrites = [], views = [];
const health = await (await fetch(origin + '/api/health')).json();
assert.equal(health.version, '1.24.7'); assert.equal(health.pid, 36552);
const browser = await chromium.launch({ channel: 'msedge', headless: true }); let result;
try {
for (const [name, viewport] of [['desktop', { width:1440, height:1000 }], ['narrow', { width:390, height:844 }]]) {
const context=await browser.newContext({viewport}), page=await context.newPage();
page.on('pageerror', error=>errors.push({name,message:error.message}));
await page.route('**/api/**',async route=>{
if (['GET','HEAD'].includes(route.request().method())) await route.continue();
else {rejectedWrites.push({name,method:route.request().method(),url:route.request().url()}); await route.abort();}
});
await page.goto(origin+'/?release=1.24.7',{waitUntil:'domcontentloaded'});
await page.locator('.connection.connected').waitFor({timeout:20000});
assert.match(await page.locator('.version-entry').textContent(),/1\.24\.7/);
const project=page.getByLabel('当前项目',{exact:true});
if(!await project.isVisible()) await page.getByRole('button',{name:'展开项目栏',exact:true}).click();
await project.selectOption('45aa4368-bd4f-4a43-8785-3853fe0c9eb1');
const panel=page.getByLabel('任务预算诊断',{exact:true});
await panel.getByRole('heading',{name:'下一次请求超出运行余额',exact:true}).waitFor({timeout:20000});
const diagnosis=await panel.innerText();
for(const pattern of [/输入可容纳/,/29,623/,/17,429/,/12,352/]) assert.match(diagnosis,pattern);
assert.equal(await page.locator('.message-input textarea').inputValue(),'');
await panel.scrollIntoViewIfNeeded();
const screenshot=path.join(root,name+'.png'); await page.screenshot({path:screenshot,fullPage:true});
fs.writeFileSync(path.join(root,name+'-diagnosis.txt'),diagnosis,{flag:'wx'});
fs.writeFileSync(path.join(root,name+'-body.txt'),await page.locator('body').innerText(),{flag:'wx'});
const overflow=await page.evaluate(()=>({viewport:innerWidth,document:document.documentElement.scrollWidth}));
assert.ok(overflow.document<=viewport.width+1);
await page.getByRole('button',{name:'打开模型中心',exact:true}).click();
const dialog=page.getByRole('dialog',{name:'模型中心',exact:true}); await dialog.waitFor();
assert.match(await dialog.innerText(),/高性能/);
const selectionScreenshot=path.join(root,name+'-model-center.png'); await page.screenshot({path:selectionScreenshot,fullPage:true});
views.push({name,viewport,screenshot,selectionScreenshot,overflow,connected:true,
originalSession:'966d83cb-d1d7-44b7-be53-9d850270e34e',originalTask:'e7ebf0d5-8ac6-47ec-bbb2-74ddefba23f4',
originalRun:'cf10d58a-d3af-4b03-9bef-77fb54837878',reservation:29623,remaining:17429,input:12352});
await context.close();
}
assert.deepEqual(errors,[]); assert.deepEqual(rejectedWrites,[]);
result={success:true,version:health.version,pid:health.pid,views,errors,rejectedWrites,realModelRequests:0,
scope:'Read-only actual production Edge render at 1440 and 390. Original R pause is displayed accurately without resuming or mutating original history. Every API write blocked; none attempted.'};
}catch(error){result={success:false,error:error.stack,views,errors,rejectedWrites,realModelRequests:0};process.exitCode=1;}
finally{await browser.close();fs.writeFileSync(path.join(root,'result.json'),JSON.stringify(result,null,2)+'\n',{flag:'wx'});}
console.log(JSON.stringify(result));
@@ -0,0 +1,50 @@
import fs from 'node:fs';
import path from 'node:path';
import assert from 'node:assert/strict';
import { chromium } from 'playwright';
const root = path.resolve('.qa/reading-reliability-s/release/browser-review-final');
assert.equal(fs.existsSync(root), false); fs.mkdirSync(root);
const origin = 'http://127.0.0.1:4318', errors = [], rejectedWrites = [], views = [];
const health = await (await fetch(origin + '/api/health')).json();
assert.equal(health.version, '1.24.7'); assert.equal(health.pid, 36552);
const browser = await chromium.launch({ channel: 'msedge', headless: true }); let result;
try {
for (const [name, viewport] of [['desktop', { width:1440, height:1000 }], ['narrow', { width:390, height:844 }]]) {
const context=await browser.newContext({viewport}), page=await context.newPage();
page.on('pageerror', error=>errors.push({name,message:error.message}));
await page.route('**/api/**',async route=>{
if (['GET','HEAD'].includes(route.request().method())) await route.continue();
else {rejectedWrites.push({name,method:route.request().method(),url:route.request().url()}); await route.abort();}
});
await page.goto(origin+'/?release=1.24.7',{waitUntil:'domcontentloaded'});
await page.locator('.connection.connected').waitFor({timeout:20000});
assert.match(await page.locator('.version-entry').textContent(),/1\.24\.7/);
const project=page.getByLabel('当前项目',{exact:true});
if(!await project.isVisible()) await page.getByRole('button',{name:'展开项目栏',exact:true}).click();
await project.selectOption('45aa4368-bd4f-4a43-8785-3853fe0c9eb1');
const panel=page.getByLabel('任务预算诊断',{exact:true});
await panel.getByRole('heading',{name:'下一次请求超出运行余额',exact:true}).waitFor({timeout:20000});
const diagnosis=await panel.innerText();
for(const pattern of [/输入可容纳/,/29,623/,/17,429/,/12,352/]) assert.match(diagnosis,pattern);
assert.equal(await page.getByRole('textbox',{name:'任务内容',exact:true}).inputValue(),'');
await panel.scrollIntoViewIfNeeded();
const screenshot=path.join(root,name+'.png'); await page.screenshot({path:screenshot,fullPage:true});
fs.writeFileSync(path.join(root,name+'-diagnosis.txt'),diagnosis,{flag:'wx'});
fs.writeFileSync(path.join(root,name+'-body.txt'),await page.locator('body').innerText(),{flag:'wx'});
const overflow=await page.evaluate(()=>({viewport:innerWidth,document:document.documentElement.scrollWidth}));
assert.ok(overflow.document<=viewport.width+1);
await page.getByRole('button',{name:'打开模型中心',exact:true}).click();
const dialog=page.getByRole('dialog',{name:'模型中心',exact:true}); await dialog.waitFor();
assert.match(await dialog.innerText(),/高性能/);
const selectionScreenshot=path.join(root,name+'-model-center.png'); await page.screenshot({path:selectionScreenshot,fullPage:true});
views.push({name,viewport,screenshot,selectionScreenshot,overflow,connected:true,
originalSession:'966d83cb-d1d7-44b7-be53-9d850270e34e',originalTask:'e7ebf0d5-8ac6-47ec-bbb2-74ddefba23f4',
originalRun:'cf10d58a-d3af-4b03-9bef-77fb54837878',reservation:29623,remaining:17429,input:12352});
await context.close();
}
assert.deepEqual(errors,[]); assert.deepEqual(rejectedWrites,[]);
result={success:true,version:health.version,pid:health.pid,views,errors,rejectedWrites,realModelRequests:0,
scope:'Read-only actual production Edge render at 1440 and 390. Original R pause is displayed accurately without resuming or mutating original history. Every API write blocked; none attempted.'};
}catch(error){result={success:false,error:error.stack,views,errors,rejectedWrites,realModelRequests:0};process.exitCode=1;}
finally{await browser.close();fs.writeFileSync(path.join(root,'result.json'),JSON.stringify(result,null,2)+'\n',{flag:'wx'});}
console.log(JSON.stringify(result));
@@ -0,0 +1,50 @@
import fs from 'node:fs';
import path from 'node:path';
import assert from 'node:assert/strict';
import { chromium } from 'playwright';
const root = path.resolve('.qa/reading-reliability-s/release/browser-review-complete');
assert.equal(fs.existsSync(root), false); fs.mkdirSync(root);
const origin = 'http://127.0.0.1:4318', errors = [], rejectedWrites = [], views = [];
const health = await (await fetch(origin + '/api/health')).json();
assert.equal(health.version, '1.24.7'); assert.equal(health.pid, 36552);
const browser = await chromium.launch({ channel: 'msedge', headless: true }); let result;
try {
for (const [name, viewport] of [['desktop', { width:1440, height:1000 }], ['narrow', { width:390, height:844 }]]) {
const context=await browser.newContext({viewport}), page=await context.newPage();
page.on('pageerror', error=>errors.push({name,message:error.message}));
await page.route('**/api/**',async route=>{
if (['GET','HEAD'].includes(route.request().method())) await route.continue();
else {rejectedWrites.push({name,method:route.request().method(),url:route.request().url()}); await route.abort();}
});
await page.goto(origin+'/?release=1.24.7',{waitUntil:'domcontentloaded'});
await page.locator('.connection.connected').waitFor({state:'attached',timeout:20000});
assert.match(await page.locator('.version-entry').textContent(),/1\.24\.7/);
const project=page.getByLabel('当前项目',{exact:true});
if(!await project.isVisible()) await page.getByRole('button',{name:'展开项目栏',exact:true}).click();
await project.selectOption('45aa4368-bd4f-4a43-8785-3853fe0c9eb1');
const panel=page.getByLabel('任务预算诊断',{exact:true});
await panel.getByRole('heading',{name:'下一次请求超出运行余额',exact:true}).waitFor({timeout:20000});
const diagnosis=await panel.innerText();
for(const pattern of [/输入可容纳/,/29,623/,/17,429/,/12,352/]) assert.match(diagnosis,pattern);
assert.equal(await page.getByRole('textbox',{name:'任务内容',exact:true}).inputValue(),'');
await panel.scrollIntoViewIfNeeded();
const screenshot=path.join(root,name+'.png'); await page.screenshot({path:screenshot,fullPage:true});
fs.writeFileSync(path.join(root,name+'-diagnosis.txt'),diagnosis,{flag:'wx'});
fs.writeFileSync(path.join(root,name+'-body.txt'),await page.locator('body').innerText(),{flag:'wx'});
const overflow=await page.evaluate(()=>({viewport:innerWidth,document:document.documentElement.scrollWidth}));
assert.ok(overflow.document<=viewport.width+1);
await page.getByRole('button',{name:'打开模型中心',exact:true}).click();
const dialog=page.getByRole('dialog',{name:'模型中心',exact:true}); await dialog.waitFor();
assert.match(await dialog.innerText(),/高性能/);
const selectionScreenshot=path.join(root,name+'-model-center.png'); await page.screenshot({path:selectionScreenshot,fullPage:true});
views.push({name,viewport,screenshot,selectionScreenshot,overflow,connected:true,
originalSession:'966d83cb-d1d7-44b7-be53-9d850270e34e',originalTask:'e7ebf0d5-8ac6-47ec-bbb2-74ddefba23f4',
originalRun:'cf10d58a-d3af-4b03-9bef-77fb54837878',reservation:29623,remaining:17429,input:12352});
await context.close();
}
assert.deepEqual(errors,[]); assert.deepEqual(rejectedWrites,[]);
result={success:true,version:health.version,pid:health.pid,views,errors,rejectedWrites,realModelRequests:0,
scope:'Read-only actual production Edge render at 1440 and 390. Original R pause is displayed accurately without resuming or mutating original history. Every API write blocked; none attempted.'};
}catch(error){result={success:false,error:error.stack,views,errors,rejectedWrites,realModelRequests:0};process.exitCode=1;}
finally{await browser.close();fs.writeFileSync(path.join(root,'result.json'),JSON.stringify(result,null,2)+'\n',{flag:'wx'});}
console.log(JSON.stringify(result));
@@ -0,0 +1,51 @@
import fs from 'node:fs';
import path from 'node:path';
import assert from 'node:assert/strict';
import { chromium } from 'playwright';
const root = path.resolve('.qa/reading-reliability-s/release/browser-review-verified');
assert.equal(fs.existsSync(root), false); fs.mkdirSync(root);
const origin = 'http://127.0.0.1:4318', errors = [], rejectedWrites = [], views = [];
const health = await (await fetch(origin + '/api/health')).json();
assert.equal(health.version, '1.24.7'); assert.equal(health.pid, 36552);
const browser = await chromium.launch({ channel: 'msedge', headless: true }); let result;
try {
for (const [name, viewport] of [['desktop', { width:1440, height:1000 }], ['narrow', { width:390, height:844 }]]) {
const context=await browser.newContext({viewport}), page=await context.newPage();
page.on('pageerror', error=>errors.push({name,message:error.message}));
await page.route('**/api/**',async route=>{
if (['GET','HEAD'].includes(route.request().method())) await route.continue();
else {rejectedWrites.push({name,method:route.request().method(),url:route.request().url()}); await route.abort();}
});
await page.goto(origin+'/?release=1.24.7',{waitUntil:'domcontentloaded'});
await page.locator('.connection.connected').waitFor({state:'attached',timeout:20000});
assert.match(await page.locator('.version-entry').textContent(),/1\.24\.7/);
const project=page.getByLabel('当前项目',{exact:true});
if(!await project.isVisible()) await page.getByRole('button',{name:'展开项目栏',exact:true}).click();
await project.selectOption('45aa4368-bd4f-4a43-8785-3853fe0c9eb1');
const panel=page.getByLabel('任务预算诊断',{exact:true});
await panel.getByRole('heading',{name:'下一次请求超出运行余额',exact:true}).waitFor({timeout:20000});
const diagnosis=await panel.innerText();
for(const pattern of [/输入可容纳/,/29,623/,/17,429/,/12,352/]) assert.match(diagnosis,pattern);
assert.equal(await page.getByRole('textbox',{name:'任务内容',exact:true}).inputValue(),'');
await panel.scrollIntoViewIfNeeded();
const screenshot=path.join(root,name+'.png'); await page.screenshot({path:screenshot,fullPage:true});
fs.writeFileSync(path.join(root,name+'-diagnosis.txt'),diagnosis,{flag:'wx'});
fs.writeFileSync(path.join(root,name+'-body.txt'),await page.locator('body').innerText(),{flag:'wx'});
const overflow=await page.evaluate(()=>({viewport:innerWidth,document:document.documentElement.scrollWidth}));
assert.ok(overflow.document<=viewport.width+1);
await page.getByRole('button',{name:'打开模型中心',exact:true}).click();
const dialog=page.getByRole('dialog',{name:'模型中心',exact:true}); await dialog.waitFor();
await dialog.getByText('高性能',{exact:true}).first().waitFor({state:'visible',timeout:15000});
assert.match(await dialog.innerText(),/高性能/);
const selectionScreenshot=path.join(root,name+'-model-center.png'); await page.screenshot({path:selectionScreenshot,fullPage:true});
views.push({name,viewport,screenshot,selectionScreenshot,overflow,connected:true,
originalSession:'966d83cb-d1d7-44b7-be53-9d850270e34e',originalTask:'e7ebf0d5-8ac6-47ec-bbb2-74ddefba23f4',
originalRun:'cf10d58a-d3af-4b03-9bef-77fb54837878',reservation:29623,remaining:17429,input:12352});
await context.close();
}
assert.deepEqual(errors,[]); assert.deepEqual(rejectedWrites,[]);
result={success:true,version:health.version,pid:health.pid,views,errors,rejectedWrites,realModelRequests:0,
scope:'Read-only actual production Edge render at 1440 and 390. Original R pause is displayed accurately without resuming or mutating original history. Every API write blocked; none attempted.'};
}catch(error){result={success:false,error:error.stack,views,errors,rejectedWrites,realModelRequests:0};process.exitCode=1;}
finally{await browser.close();fs.writeFileSync(path.join(root,'result.json'),JSON.stringify(result,null,2)+'\n',{flag:'wx'});}
console.log(JSON.stringify(result));
@@ -0,0 +1,58 @@
import fs from 'node:fs';
import path from 'node:path';
import assert from 'node:assert/strict';
import { spawn } from 'node:child_process';
const root = '.qa/reading-reliability-s/release';
const read = file => JSON.parse(fs.readFileSync(file, 'utf8').replace(/^\uFEFF/, ''));
const mode = process.argv[2];
const freeze = read(root + '/product-freeze.json'), release = read('releases/latest.json');
assert.equal(release.documentation.sourceCommit, freeze.sourceCommit);
assert.equal(read(path.join(release.portableRoot, 'MANIFEST.json')).version, '1.24.7');
if (mode === 'register') {
fs.writeFileSync(root + '/candidate.json', JSON.stringify(release, null, 2) + '\n', { flag: 'wx' });
console.log(JSON.stringify({ candidate: release.portableRoot, sourceCommit: freeze.sourceCommit }));
process.exit(0);
}
assert.deepEqual(release, read(root + '/candidate.json'));
const commands = {
rebuild: [process.execPath, ['scripts/verify-packaged-build.mjs']],
archives: ['C:/Users/wangy/AppData/Local/Programs/Python/Python312/python.exe', ['-I', 'scripts/verify-archives.py']],
portable: [process.execPath, ['scripts/verify-release.mjs']],
'portable-recheck': [process.execPath, ['scripts/verify-release.mjs']],
'portable-ui': [process.execPath, ['scripts/verify-reading-reliability-ui.mjs']],
install: [process.execPath, ['.qa/reading-s-install-lifecycle.mjs']]
};
assert.ok(commands[mode]);
const output = root + '/' + mode + '.json';
assert.equal(fs.existsSync(output), false, 'Do not replay unknown or failed result.');
const log = root + '/' + mode + '.log', fd = fs.openSync(log, 'wx');
const env = { ...process.env, AGENT_WORKBENCH_TEST_PYTHON: 'C:/Users/wangy/AppData/Local/Programs/Python/Python312/python.exe',
...(['portable-ui', 'install'].includes(mode) ? { AGENT_WORKBENCH_QA_APP_ROOT: release.portableRoot } : {}) };
let code, report; const startedAt = new Date().toISOString();
try {
console.log('START ' + mode);
code = await new Promise((resolve, reject) => {
const child = spawn(...commands[mode], { env, windowsHide: true, stdio: ['ignore', fd, fd] });
child.once('error', reject); child.once('close', resolve);
});
fs.closeSync(fd);
assert.equal(code, 0, 'Inspect preserved log ' + log);
assert.deepEqual(read('releases/latest.json'), release);
const lines = fs.readFileSync(log, 'utf8').trim().split(/\r?\n/);
let resultPath;
if (mode === 'archives') resultPath = '.qa/archives-validation.json';
else if (['portable', 'portable-recheck'].includes(mode)) resultPath = '.qa/release-validation.json';
else if (mode === 'install') resultPath = read('.qa/desktop-install-location.json').result;
else {
const summary = JSON.parse(lines.findLast(line => line.startsWith('{"success":')));
resultPath = summary.output ?? path.join(summary.root, 'result.json');
}
const result = read(resultPath); assert.equal(result.success, true);
if (mode === 'rebuild') assert.equal(result.sourceCommit, freeze.sourceCommit);
if (mode === 'install') assert.equal(fs.realpathSync.native(result.packagedSource), fs.realpathSync.native(release.portableRoot));
report = { success: true, resultPath, result };
} catch (error) { report = { success: false, code, error: error.stack }; process.exitCode = 1; }
report = { ...report, mode, startedAt, finishedAt: new Date().toISOString(), log,
sourceCommit: freeze.sourceCommit, candidate: release.portableRoot, realModelRequests: 0 };
fs.writeFileSync(output, JSON.stringify(report, null, 2) + '\n', { flag: 'wx' });
console.log(JSON.stringify({ success: report.success, mode, output, resultPath: report.resultPath, error: report.error }));
@@ -0,0 +1,20 @@
import fs from 'node:fs';
import path from 'node:path';
import assert from 'node:assert/strict';
import { execFileSync } from 'node:child_process';
import { createHash } from 'node:crypto';
const read = file => JSON.parse(fs.readFileSync(file, 'utf8'));
const hash = file => createHash('sha256').update(fs.readFileSync(file)).digest('hex');
const root = '.qa/reading-reliability-s/release';
const freeze = read(root + '/product-freeze.json'), build = read(root + '/rebuild.json').result;
assert.equal(build.success, true);
const extra = build.inputs.filter(row => !freeze.files.some(input => input.path === row.path));
assert.deepEqual(extra.map(row => row.path).sort(), ['cli.mjs','normalize.mjs','package.json','render.mjs','summary.mjs'].map(name => 'examples/delivery-report/dist/' + name).sort());
for (const input of freeze.files) assert.equal(build.inputs.find(row => row.path === input.path)?.sha256, input.sha256);
execFileSync(process.execPath, ['scripts/build.mjs'], { cwd: path.join(build.destination, 'examples/delivery-report'), windowsHide: true });
for (const input of extra) assert.equal(hash(path.join(build.destination, input.path)), input.sha256);
const result = { success: true, trackedInputs: freeze.files.length, archiveInputs: build.inputCount, extra, unchanged: true,
sourceCommit: freeze.sourceCommit, realModelRequests: 0,
scope: 'Five declared ignored example dist inputs rebuilt independently from exact extracted source ZIP and matched byte hashes. No immutable payload changes.' };
fs.writeFileSync(root + '/extra-example-inputs.json', JSON.stringify(result, null, 2) + '\n', { flag: 'wx' });
console.log(JSON.stringify(result));
@@ -0,0 +1,59 @@
import fs from 'node:fs';
import assert from 'node:assert/strict';
const read=file=>JSON.parse(fs.readFileSync(file,'utf8'));
const root='.qa/reading-reliability-s', final=read(root+'/formal-release/final.json');assert.equal(final.success,true);
const manifest=read('docs/reading-s-evidence/manifest.json'),identity=read('docs/reading-s-release-identity.json');
const offline=read('docs/reading-s-offline.txt');assert.equal(offline.success,true);
const timing=read(root+'/release/update-timing.json');
const short=identity.sourceCommit.slice(0,8),candidate=identity.candidate.split('-').at(-1);
const summary=`S已本地交付1.24.7/source${short}/候选${candidate}:文件续读携带原范围/偏移/SHA、纯文件批次容量及未用份额分配、Run/Task预留余额准确解释。96相关和18安装升级、源码/便携/实际安装新UI各10组、20组历史UI、565输入独立重建22/22、三归档/便携及12实际安装和16+9文件流程通过。正式原PID36552/Provider、63项目79会话6262记录13279事件及outbox/3660工件,所有原表行/全导出/工件与公开模型中心在网页/原生回看后再次精确保持,9229关闭。${manifest.count}份/${manifest.bytes}字节证据离线通过,见docs/QA-READING-RELIABILITY-S.md及docs/reading-s-release-identity.json。原R三响应+固定续读回放由8轮未完到6轮读完2584;原九测试仍失败,0修改/交付,本轮真实模型请求0。下一按docs/LIVE-DELIVERY-T-PLAN.md另冻结单次预检/新Node,16K/2048/8请求/90000tokens;不续跑旧失败,Node通过再React/Python/Purge。更新${timing.operationMs}ms重复核验及1.25独立排期。不委派、不提取/复制/重存Key、不重启原Provider、不推送。`;
const write=(file,body)=>fs.writeFileSync(file,body,{flag:'wx'});
const edit=(file,before,after)=>{const body=fs.readFileSync(file,'utf8');assert.equal(body.split(before).length,2,file);fs.writeFileSync(file,body.replace(before,after));};
const prepend=(file,body)=>{const original=fs.readFileSync(file,'utf8'),end=original.indexOf('\n');fs.writeFileSync(file,original.slice(0,end+1)+'\n'+body+'\n\n'+original.slice(end+1).replace(/^\s*\n/,''));};
write('docs/QA-READING-RELIABILITY-S.md',`# S:读取可靠性与预算暂停解释验收
1.24.7 已本地交付,source \`${identity.sourceCommit}\`,不可变候选 \`${identity.candidate}\`。固定发行身份见 [reading-s-release-identity.json](reading-s-release-identity.json)。正式交付完成时间以原 [最终收据](reading-s-evidence/receipts/formal-release/final.json) 为准;包内“发行验收中”文档是冻结时点快照,未为补写结论改包。
## 本次行为
文件续读明确区分恢复已返回页与读取未返回的原文件,携带原路径、行范围、偏移和可选 SHA-256;版本变化拒绝拼接。纯文件批次按完整转义消息预算分配并回收同批未用空间,混合操作与工件读取保留原保守预留。旧发送视图保持可核验。下一请求预留超出 Run/Task 余额时,分别展示所需与剩余;原预算、1600 安全预留、目标/约束、观察用量和上游原文保持。详见 [实现说明](READING-RELIABILITY-S-IMPLEMENTATION.md)。
## 已完成验证
- 96/96 相关测试及 18/18 安装/旧版升级回归通过;TypeScript/Vite、移动、服务与桌面构建通过。
- 源码、便携和实际安装副本的新暂停界面各 10 组通过;1440px/390px 四类诊断、刷新、续接预览及取消保持原 Task/Run,未发出模型请求。历史上下文 11、预算 2、模型选择 7 组通过;390px 是浏览器视口,本轮未重新做 iPhone 实机联调。
- 三协议实际请求验证 Unicode/emoji、CRLF、长行和小/大源码页、历史版本及实际发送容量。R 原前三响应加固定续读策略:旧实现 8 请求仍只在后续请求可见到 2173/2584,新实现 6 请求完整读取 2584。原九测试仍全部失败、build 通过,0 源码修改或 Git 交付;这是离线夹具改善,不是真实模型成功或 Purge 收益。
- 源码 ZIP 565 输入独立离线重建,22/22 编译输出与不可变便携一致;560 跟踪输入逐 SHA 对照,另 5 份示例 dist 用原 ZIP 源码/脚本独立重建匹配。三归档 CRC、清单、目录、字节与 SHA 核对通过;实际便携启动、REA/受控引擎/浏览流程通过。
- 12 组实际隔离安装、窗口/请求/审批保护、清单故障、指针写入后回退、正常更新、原 PID/凭据保留、重启和卸载通过;实际安装副本的成果 16 组、文件/HTML 9 组及 S 新 UI 10 组通过。测试只用自有本机 HTTP 和凭据夹具。
## 正式升级与回看
实际 Setup 将 1.24.6 升级到 1.24.7,保留原 PID36552 和 Provider。63 项目、79 会话、6262 记录、13279 事件及 outbox、3660 工件文件的全部原表行、全导出、原工件字节和公开模型中心精确保持;默认高性能池、三条原连接/原修订和旧 apikey 停用状态保持。没有读取、复制或重存 Key,没有重启原 Provider,没有恢复旧 Task。
正式网页 1440px/390px 阻断全部 API 写入,实际旧 R 暂停显示输入 12352 可容纳、预留 29623、余额 17429,未改写历史或续跑。正式原生窗口经原安装入口打开后回看;其后再次核对全部原状态,9229 关闭。原生原图格式与文件扩展名已核对。
本次安装操作 ${timing.operationMs} ms,驱动总耗时 ${timing.totalDriverMs} ms。完整清单与目录重复核验保持,详见 [耗时](reading-s-evidence/receipts/release/update-timing.json),后续性能优化独立验证。
## 原失败与证据
首计划 ID/混合视图判据、读取分配与历史版本回归失败、UI 用量字段错误及修正全部保留。首便携启动被 HOST_PORT_UNVERIFIED 拦下;确认隔离目录无启动产物、无所属宿主后,另建隔离样本完整通过,未改候选或降低保护,唯一原因仍未知。取证助手首清单字段错误、正式网页验收器旧定位/隐藏栏/弹窗加载等待失败,以及原生最小化/用户输入拒绝和启动窗口迟到同样保留;后续修正定位/等待并用真实回看确认,没有重装或改写正式记录。
[证据清单](reading-s-evidence/manifest.json)含 ${manifest.count} 份、${manifest.bytes} 字节,SHA-256 \`${offline.manifestSha256}\`。新旧源码、原 R 输入引用、请求/计划回放、首失败、最终测试、独立源码/发行/安装、正式状态和原图按原字节封存;旧源码 node_modules 连接未递归复制。
[离线核验器](../scripts/verify-reading-s-evidence.mjs)逐文件及语义核对通过,见 [收据](reading-s-offline.txt)。加 --index 核对 Git 暂存原字节,收据为 [reading-s-git-index.txt](reading-s-git-index.txt)。本地范围提交,不推送。阶段取舍见 [回顾](RETROSPECTIVE-READING-S.md),下一按 [T 新真实样本计划](LIVE-DELIVERY-T-PLAN.md) 执行;本轮真实模型请求 0。
`);
write('docs/RETROSPECTIVE-READING-S.md',fs.readFileSync('.qa/reading-s-retro-draft.md'));
write('docs/LIVE-DELIVERY-T-PLAN.md',fs.readFileSync('.qa/reading-s-next-plan-draft.md'));
prepend('AGENTS.md',summary);
edit('AGENTS.md','正式服务入口http://127.0.0.1:4318,当前1.24.6,沿用原PID36552与Provider;不可变程序身份以docs/network-q-release-identity.json为准,R后配置/数据与实测见docs/QA-LIVE-DELIVERY-R.md。R已按声明核对原全表/导出/工件保留及原预算/池恢复;普通源码/规划/Git工作不重启服务、不读取或复制Key。正式原生窗口按Q实际回看记录,R没有新增原生/UI验收结论,旧F的Escape失败作为历史保留。',
'正式服务入口http://127.0.0.1:4318,当前1.24.7,沿用原PID36552与Provider;不可变程序身份以docs/reading-s-release-identity.json为准,S正式保留状态与网页/原生回看见docs/QA-READING-RELIABILITY-S.md,最近真实模型仍是R未交付原样本。S核对原全表/导出/工件/模型中心精确保持,普通源码/规划/Git工作不重启服务、不读取或复制Key。下一按docs/LIVE-DELIVERY-T-PLAN.md另冻结新样本,旧F的Escape失败保持历史含义。');
prepend('docs/NEXT-STEPS.md',summary+'\n\nS已完成原计划。下一优先[T新真实样本](LIVE-DELIVERY-T-PLAN.md),先单次预检、同预算新Node,再依前置条件扩React/Python及原版PurgeAB/BA。下方R及更早段落按其历史时点理解。');
prepend('docs/VALIDATION.md',summary);
edit('CHANGELOG.md','## 1.24.7 · 文件续读与预算解释(发行验收中)','## 1.24.7 · 文件续读与预算解释(已本地交付)');
edit('docs/READING-RELIABILITY-S-IMPLEMENTATION.md','本切片接续 R 原失败,产品版本 1.24.7,发行验收中。正式升级前仍以 1.24.6 为准。后续发行身份、保留状态与原始检查结果另见最终 QA,不用开发回放代替上线或真实模型交付。',
'本切片接续 R 原失败,1.24.7 已完成本地发行及保留原 PID/Provider 升级。发行身份、状态保留与原始结果见 [最终 QA](QA-READING-RELIABILITY-S.md);开发回放与真实模型交付分别判定,候选包内该文档保留冻结时点原文。');
edit('task_plan.md','5. [in_progress] 冻结1.24.7不可变新候选,独立源码输入/重建/归档/便携/隔离安装和正式保留PID/Provider升级;原数据核对、实际入口/原生回看、证据封存/离线核验/回顾与Git收尾。',
'5. [complete] 1.24.7/source'+short+'不可变候选独立565输入/22编译、三归档/便携、12实际安装及16+9+10安装界面通过;正式保留原PID36552/Provider及全部原表行/导出/工件/模型中心,网页/原生回看后再次相同。证据逐字节离线通过,回顾与新T计划已记录;本地范围Git收尾,不推送。');
prepend('task_plan.md','S发行完成:'+summary);
for(const file of['findings.md','progress.md'])prepend(file,'## S发行与回顾完成\n\n'+summary);
console.log(JSON.stringify({success:true,version:identity.version,sourceCommit:identity.sourceCommit,evidenceCount:manifest.count}));
@@ -0,0 +1,12 @@
import fs from 'node:fs';
const folder='C:/Users/wangy/AppData/Local/Programs/AgentWorkbench/.installation';
const records=fs.readdirSync(folder).filter(name=>/^progress-.*\.json$/.test(name)).map(name=>JSON.parse(fs.readFileSync(folder+'/'+name))).sort((a,b)=>b.updatedAt.localeCompare(a.updatedAt));
const r=records[0];
console.log(JSON.stringify({state:r.state,stage:r.stage,scope:r.scope,counts:r.counts,updatedAt:r.updatedAt,elapsedMs:r.elapsedMs,candidate:r.candidate?.directory}));
const log='.qa/reading-reliability-s/formal-release/installer.log';
if(fs.existsSync(log)){
const size=fs.statSync(log).size,fd=fs.openSync(log,'r'),b=Buffer.alloc(Math.min(size,6000));
fs.readSync(fd,b,0,b.length,size-b.length);fs.closeSync(fd);
const progress=b.toString('utf8').split(/\r?\n/).filter(line=>/AWBPROGRESS|AWBFILES|Log closed|Need to restart/.test(line));
console.log(progress.slice(-2).join('\n'));
}
@@ -0,0 +1,85 @@
import fs from 'node:fs';
import path from 'node:path';
import assert from 'node:assert/strict';
import net from 'node:net';
import { spawn, execFileSync } from 'node:child_process';
import { publicSnapshot, compareOriginalState, hash } from './desktop-f-state.mjs';
import { inspectHostInstallation } from '../src/core/desktop/app-installation.mjs';
const root = path.resolve('.qa/reading-reliability-s/formal-release'); fs.mkdirSync(root, { recursive: true });
const read = file => JSON.parse(fs.readFileSync(file, 'utf8'));
const save = (name, value) => fs.writeFileSync(path.join(root, name), JSON.stringify(value, null, 2) + '\n', { flag: 'wx' });
const release = read('releases/latest.json'), sourceCommit = release.documentation.sourceCommit;
const manifest = read(path.join(release.portableRoot, 'MANIFEST.json')); assert.equal(manifest.version, '1.24.7');
const url = 'http://127.0.0.1:4318', installRoot = 'C:/Users/wangy/AppData/Local/Programs/AgentWorkbench';
const dataRoot = 'C:/Users/wangy/AppData/Local/Packages/OpenAI.Codex_2p2nqsd0c76g0/LocalCache/Local/AgentWorkbench';
const mode = process.argv[2]; assert.ok(['before', 'install', 'final'].includes(mode));
assert.ok(!fs.existsSync(path.join(root, mode + '.json')), 'Keep prior receipts; do not retry unknown installation.');
const startedAt = new Date().toISOString();
async function api(route) {
const boot = await (await fetch(url + '/api/bootstrap', { signal: AbortSignal.timeout(20000) })).json();
const response = await fetch(url + '/api' + route, { headers: { 'X-Workbench-Token': boot.token, Connection: 'close' }, signal: AbortSignal.timeout(20000) });
assert.equal(response.status, 200, route); return response.json();
}
async function inspectorClosed() {
await new Promise((resolve, reject) => {
const socket = net.createConnection({ host: '127.0.0.1', port: 9229 }); socket.setTimeout(1500);
socket.once('connect', () => { socket.destroy(); reject(Error('Inspector open.')); });
socket.once('error', error => { socket.destroy(); error.code === 'ECONNREFUSED' ? resolve() : reject(error); });
socket.once('timeout', () => { socket.destroy(); reject(Error('Inspector state unknown.')); });
});
}
let result;
try {
assert.equal(execFileSync('git', ['rev-parse', 'HEAD'], { encoding: 'utf8' }).trim(), sourceCommit);
assert.equal(execFileSync('git', ['status', '--porcelain'], { encoding: 'utf8' }).trim(), '');
await inspectorClosed();
const current = await publicSnapshot(url, dataRoot), host = await api('/host/status');
assert.equal(current.pid, 36552); assert.equal(current.hasKey, true); assert.deepEqual(current.active, []); assert.equal(host.activity.active, false);
if (mode === 'before') {
compareOriginalState(read('.qa/reading-reliability-s/production-before.json'), current);
assert.equal(current.version, '1.24.6'); assert.equal(host.desktopWindows.count, 0);
assert.match(current.runtimeRoot.replaceAll('\\', '/'), /\/versions\/Agent-Workbench-1\.24\.6-win-x64-20261103151145294$/);
assert.equal(read('.qa/reading-reliability-s/release/preinstall-review.json').success, true);
const archived = read('.qa/archives-validation.json'), portable = read('.qa/release-validation.json');
assert.equal(archived.success, true); assert.equal(portable.success, true); assert.equal(portable.appRoot, release.portableRoot);
save('private-before.json', current); save('pointer-before.json', read(path.join(installRoot, 'current.json'))); save('release.json', release);
result = { success: true, version: current.version, pid: current.pid, projects: current.projects,
sessions: current.database.sessionIds.length, tables: Object.fromEntries(Object.entries(current.database.tables).map(([name, table]) => [name, table.count])),
runtimeRoot: current.runtimeRoot, sourceCommit, hasKey: current.hasKey, active: current.active };
} else {
const before = read(path.join(root, 'private-before.json')); compareOriginalState(before, current);
assert.deepEqual(release, read(path.join(root, 'release.json')));
if (mode === 'install') {
assert.equal(current.version, '1.24.6'); assert.equal(host.desktopWindows.count, 0);
assert.deepEqual(read(path.join(installRoot, 'current.json')), read(path.join(root, 'pointer-before.json')));
assert.equal(hash(fs.readFileSync(release.installer.path)), release.installer.sha256);
const logFile = path.join(root, 'installer.log'), fd = fs.openSync(path.join(root, 'installer-process.log'), 'wx'); let code;
try {
code = await new Promise((resolve, reject) => {
const child = spawn(release.installer.path, ['/VERYSILENT', '/SUPPRESSMSGBOXES', '/NORESTART', '/LOG=' + logFile,
'/DIR=' + installRoot, '/DATA_ROOT=' + dataRoot, '/PORT=4318'], { windowsHide: true, stdio: ['ignore', fd, fd] });
child.once('error', reject); child.once('close', resolve);
});
} finally { fs.closeSync(fd); }
save('installer-exit.json', { code, finishedAt: new Date().toISOString() }); assert.equal(code, 0, 'Unknown/failed install is not automatically replayed.');
} else assert.equal(read(path.join(root, 'install.json')).success, true);
const after = await publicSnapshot(url, dataRoot), retention = compareOriginalState(before, after);
assert.equal(after.version, '1.24.7'); assert.equal(after.pid, before.pid); assert.equal(after.hasKey, true);
const candidate = path.join(installRoot, 'versions', path.basename(release.portableRoot));
assert.equal(fs.realpathSync.native(after.runtimeRoot), fs.realpathSync.native(candidate));
const identity = inspectHostInstallation(candidate), pointer = read(path.join(installRoot, 'current.json'));
assert.equal(pointer.active.directory, path.basename(candidate)); assert.equal(pointer.active.manifestSha256, identity.manifestSha256);
assert.equal(pointer.active.runtimeIdentity, identity.identitySha256);
const operation = fs.readdirSync(path.join(installRoot, '.installation')).filter(name => /^progress-.*\.json$/.test(name))
.map(name => read(path.join(installRoot, '.installation', name))).sort((a, b) => b.updatedAt.localeCompare(a.updatedAt))[0];
assert.equal(operation.state, 'completed'); assert.equal(operation.candidate.manifestSha256, identity.manifestSha256);
save(mode + '-private-after.json', after);
result = { success: true, version: after.version, pid: after.pid, runtimeRoot: after.runtimeRoot, projects: after.projects,
sessions: after.database.sessionIds.length, hasKey: after.hasKey, retention, pointer, operation,
desktopWindows: (await api('/host/status')).desktopWindows.count, inspectorClosed: true, sourceCommit };
}
} catch (error) { result = { success: false, error: error.stack }; process.exitCode = 1; }
result = { ...result, mode, startedAt, finishedAt: new Date().toISOString(), realModelRequests: 0,
scope: 'Public API/full original row/export/artifact hashes; installer retains original running host and Provider. No credential extraction or task replay.' };
save(mode + '.json', result); console.log(JSON.stringify(result));
@@ -0,0 +1,15 @@
import fs from 'node:fs';
import path from 'node:path';
import assert from 'node:assert/strict';
import { execFileSync } from 'node:child_process';
import { createHash } from 'node:crypto';
const root = '.qa/reading-reliability-s/release'; fs.mkdirSync(root, { recursive: true });
assert.equal(execFileSync('git', ['status', '--porcelain'], { encoding: 'utf8' }).trim(), '');
const names = execFileSync('git', ['ls-files', '-z', '--', 'src', 'public', 'scripts', 'tests', 'examples', 'benchmarks', 'research', 'package.json', 'package-lock.json', 'vite.config.ts', 'tsconfig.json', 'index.html', 'mobile.html'], { encoding: 'utf8' }).split('\0').filter(Boolean);
const record = file => { const bytes = fs.readFileSync(file); return { path: file.replaceAll('\\', '/'), bytes: bytes.length, sha256: createHash('sha256').update(bytes).digest('hex') }; };
const walk = folder => fs.readdirSync(folder, { withFileTypes: true }).flatMap(entry => entry.isDirectory() ? walk(path.join(folder, entry.name)) : [path.join(folder, entry.name)]);
const value = { sourceCommit: execFileSync('git', ['rev-parse', 'HEAD'], { encoding: 'utf8' }).trim(), version: '1.24.7',
files: names.sort().map(record), compiled: ['dist-client', 'dist-server', 'dist-desktop'].flatMap(walk).sort().map(record),
createdAt: new Date().toISOString(), realModelRequests: 0 };
fs.writeFileSync(path.join(root, 'product-freeze.json'), JSON.stringify(value, null, 2) + '\n', { flag: 'wx' });
console.log(JSON.stringify({ sourceCommit: value.sourceCommit, sourceInputs: value.files.length, compiled: value.compiled.length }));
@@ -0,0 +1,182 @@
import assert from 'node:assert/strict';
import fs from 'node:fs';
import path from 'node:path';
import net from 'node:net';
import http from 'node:http';
import { spawn, execFile } from 'node:child_process';
import { promisify } from 'node:util';
import { createHash, randomUUID } from 'node:crypto';
import { _electron } from 'playwright';
import { planDesktopBundle, copyDesktopBundle, buildNativeLauncher } from '../scripts/bundle-desktop.mjs';
import { buildWindowsInstaller } from '../scripts/build-installer.mjs';
import { HostController } from '../src/core/desktop/host-controller.mjs';
import { InstallationManager, snapshotPersistentData } from '../src/core/desktop/install-manager.mjs';
import { observeWindowsProcess } from '../src/core/desktop/process-identity.mjs';
import { projectEnvironment } from '../src/core/process-environment.mjs';
if (process.argv.includes('--help')) { console.log('node scripts/verify-desktop-install.mjs — isolated actual user-level setup/update/uninstall, full data retention, owned DPAPI/HTTP fixture, active-window and request guards, fault rollback and original PID retention. No configured-model credentials.'); process.exit(0); }
assert.equal(process.argv.length, 2);
const appRoot = process.cwd(), root = path.resolve('.qa', 'desktop-install-' + randomUUID()), installRoot = path.join(root, 'program'), dataRoot = path.join(root, 'user-data');
const packagedSource = process.env.AGENT_WORKBENCH_QA_APP_ROOT ? fs.realpathSync.native(process.env.AGENT_WORKBENCH_QA_APP_ROOT) : null;
const version = JSON.parse(fs.readFileSync('package.json')).version, results = [], appId = randomUUID(), hash = value => createHash('sha256').update(value).digest('hex');
fs.mkdirSync(root, { recursive: true });
const reserved = net.createServer(); await new Promise(resolve => reserved.listen(0, '127.0.0.1', resolve)); const port = reserved.address().port; await new Promise(resolve => reserved.close(resolve));
const record = (name, detail = {}) => { results.push({ name, passed: true, ...detail }); console.log('PASS ' + name); };
const until = async (action, timeout = 60000) => { const end = Date.now() + timeout; while (!await action()) { if (Date.now() > end) throw Error('Timeout: ' + stage); await new Promise(resolve => setTimeout(resolve, 100)); } };
const run = promisify(execFile), env = projectEnvironment();
let stage = 'prepare owned full releases', controller, application, ownedModel, report, heldRequest, releaseRequest, original, installedUI = null;
const launcher = path.resolve('.qa/native-launcher/AgentWorkbench.exe'); buildNativeLauncher(appRoot, launcher); const desktop = planDesktopBundle(appRoot);
function payload(name) {
const directory = path.join(root, name); fs.mkdirSync(directory);
if (packagedSource) {
const original = JSON.parse(fs.readFileSync(path.join(packagedSource, 'MANIFEST.json')));
assert.equal(original.kind, 'portable'); assert.equal(original.version, version);
for (const entry of original.files) { const file = path.join(packagedSource, entry.path); assert.equal(hash(fs.readFileSync(file)), entry.sha256, entry.path); }
fs.cpSync(packagedSource, directory, { recursive: true }); fs.unlinkSync(path.join(directory, 'MANIFEST.json'));
} else {
for (const item of ['dist-client', 'dist-server', 'dist-desktop', 'public', 'examples']) fs.cpSync(path.join(appRoot, item), path.join(directory, item), { recursive: true });
fs.mkdirSync(path.join(directory, 'scripts')); fs.copyFileSync('scripts/python-check-runner.py', path.join(directory, 'scripts/python-check-runner.py'));
fs.mkdirSync(path.join(directory, 'runtime')); fs.copyFileSync(process.execPath, path.join(directory, 'runtime/node.exe'));
fs.cpSync(path.join(path.dirname(process.execPath), 'node_modules/npm'), path.join(directory, 'runtime/node_modules/npm'), { recursive: true });
copyDesktopBundle(appRoot, directory, desktop, { launcher });
}
fs.writeFileSync(path.join(directory, 'owned-release.txt'), name + '\n');
// A valid PE overlay explicitly models a future owned native bootstrap with
// different bytes. The updated root entry must follow it and roll back with it.
if (name.startsWith('owned-b-')) fs.appendFileSync(path.join(directory, 'AgentWorkbench.exe'), '\nowned-native-bootstrap-update-fixture\n');
const walk = relative => fs.readdirSync(path.join(directory, relative), { withFileTypes: true }).flatMap(entry => {
const next = path.posix.join(relative, entry.name), file = path.join(directory, next); return entry.isDirectory() ? walk(next) : [{ path: next, bytes: fs.statSync(file).size, sha256: hash(fs.readFileSync(file)) }];
});
fs.writeFileSync(path.join(directory, 'MANIFEST.json'), JSON.stringify({ app: 'agent-workbench', kind: 'portable', version, platform: 'win32-x64', nodeVersion: process.version, files: walk('') }, null, 2)); return directory;
}
async function setup(installer, label, expected = 0) {
const log = path.join(root, label + '-setup.log');
let code = 0, output = '';
try { const result = await run(installer.path, ['/VERYSILENT', '/SUPPRESSMSGBOXES', '/NORESTART', '/LOG=' + log, '/DIR=' + installRoot, '/DATA_ROOT=' + dataRoot, '/PORT=' + port], { windowsHide: true, timeout: 300000, maxBuffer: 1024 * 1024, env }); output = result.stdout + result.stderr; }
catch (error) { code = typeof error.code === 'number' ? error.code : -1; output = String(error.stdout ?? '') + String(error.stderr ?? ''); }
fs.writeFileSync(path.join(root, label + '-process.txt'), output);
if (expected === 0) assert.equal(code, 0, 'Actual installer failed; inspect ' + log); else assert.notEqual(code, 0, 'Actual installer unexpectedly accepted the rejected condition.');
const logText = fs.readFileSync(log, 'utf8'); assert.match(logText, /AWBPROGRESS /, 'Actual installer must receive original stage output.');
const records = fs.readdirSync(path.join(installRoot, '.installation')).filter(name => /^progress-.*\.json$/.test(name)).map(name => JSON.parse(fs.readFileSync(path.join(installRoot, '.installation', name)))).sort((a, b) => b.updatedAt.localeCompare(a.updatedAt));
assert.equal(records[0].state, expected === 0 ? 'completed' : 'failed');
if (expected === 0) { assert.ok(records[0].history.some(row => row.stage === 'verify-candidate' && row.scope === 'full-manifest' && row.counts.completedFiles === row.counts.totalFiles)); assert.match(logText, /AWBFILES \d+\/\d+/); }
return { code, log, sha256: hash(fs.readFileSync(log)) };
}
async function launchElectron() {
const current = JSON.parse(fs.readFileSync(path.join(installRoot, 'current.json'))), directory = path.join(installRoot, 'versions', current.active.directory);
const client = await _electron.launch({ executablePath: path.join(directory, 'desktop/AgentWorkbench.exe'), args: ['--app-root', directory, '--data-root', dataRoot, '--port', String(port), '--hidden-window'], cwd: directory, env, timeout: 60000 });
const page = await client.firstWindow(); await page.locator('.connection.connected').waitFor({ timeout: 30000 }); return client;
}
function ownedManager(candidateRoot) { return new InstallationManager({ installRoot, dataRoot, candidateRoot, port }); }
try {
const a = payload('owned-a-' + version), b = payload('owned-b-' + version), fault = payload('owned-fault-' + version);
stage = 'compile isolated actual installers';
const firstSetup = buildWindowsInstaller({ root: appRoot, portableRoot: a, outputRoot: root, qaFixture: { appId } });
const secondSetup = buildWindowsInstaller({ root: appRoot, portableRoot: b, outputRoot: root, qaFixture: { appId } });
const faultSetup = buildWindowsInstaller({ root: appRoot, portableRoot: fault, outputRoot: root, qaFixture: { appId, fault: 'corrupt-installed-candidate' } });
stage = 'actual clean user-level installation'; const installed = await setup(firstSetup, 'clean');
const activeA = path.join(installRoot, 'versions', path.basename(a)); controller = new HostController({ appRoot: activeA, dataRoot, port, nodePath: process.execPath });
original = (await controller.start()).identity; application = await launchElectron();
assert.equal((await controller.status()).desktopWindows.count, 1); record(stage, { installer: firstSetup, ...installed, hostPid: original.pid, electronPid: await application.evaluate(() => process.pid) });
stage = 'actual installer rejects an open attested desktop window'; const windowRejection = await setup(secondSetup, 'window-block', 1);
assert.equal(ownedManager(null).current().active.directory, path.basename(a)); assert.equal((await controller.connect()).identity.pid, original.pid); record(stage, windowRejection);
await application.close(); application = null; await until(async () => (await controller.status()).desktopWindows.count === 0);
const key = 'owned-install-credential-' + randomUUID(); let authenticated = 0, hold = false;
ownedModel = http.createServer(async (request, response) => {
const chunks = []; for await (const bytes of request) chunks.push(bytes);
assert.equal(request.headers.authorization, 'Bearer ' + key); authenticated++;
if (hold) { heldRequest = true; await new Promise(resolve => { releaseRequest = resolve; request.on('close', resolve); }); }
response.writeHead(200, { 'Content-Type': 'application/json' }); response.end(JSON.stringify({ choices: [{ message: { role: 'assistant', content: 'OK' }, finish_reason: 'stop' }], usage: { prompt_tokens: 4, completion_tokens: 1, total_tokens: 5 } }));
}); await new Promise(resolve => ownedModel.listen(0, '127.0.0.1', resolve));
const connection = await controller.api('/model-center/connections', { method: 'POST', input: { name: 'Owned desktop install credential', model: 'owned', baseUrl: 'http://127.0.0.1:' + ownedModel.address().port + '/v1', apiKey: key, persistKey: true } });
const connectionId = connection.id ?? connection.connection?.id ?? connection.center?.connections.at(-1).id;
assert.ok(connectionId); await controller.api('/model-center/probe', { method: 'POST', input: { id: connectionId } }); assert.equal(authenticated, 1);
stage = 'actual update rejects a pending owned model request'; hold = true;
// Manifest verification is synchronous and can occupy this fixture's event
// loop. Keep the explicitly held local response pending across that work;
// this is the QA client's deadline, not a product/model timeout override.
const pendingStarted = Date.now();
const pending = controller.api('/model-center/probe', { method: 'POST', input: { id: connectionId }, timeoutMs: 300000 }); pending.catch(() => {});
await until(() => heldRequest); await assert.rejects(ownedManager(null).preflight(), /任务|请求/);
assert.equal((await controller.connect()).identity.pid, original.pid);
assert.equal(authenticated, 2, 'Exactly the initial and explicitly held owned requests; no retry.');
record(stage, { fixture: 'Actual owned HTTP response held until full manifest/preflight and original PID verification settle. No configured-model call or product timeout change.', elapsedMs: Date.now() - pendingStarted });
releaseRequest?.(); hold = false; assert.equal((await pending).ok, true);
stage = 'actual installer rejects a persistent original task approval';
const projectRoot = path.join(root, 'approval-project'); fs.mkdirSync(projectRoot); fs.writeFileSync(path.join(projectRoot, 'README.md'), '# Owned installer approval project\n');
const project = await controller.api('/projects', { method: 'POST', input: { root: projectRoot } }), session = await controller.api('/sessions', { method: 'POST', input: { projectId: project.id } });
await controller.api('/sessions/' + session.id + '/messages', { method: 'POST', input: { content: 'Read this owned project and prepare the offline demonstration proposal.' } });
await until(async () => (await controller.api('/sessions/' + session.id)).approvals.some(value => value.status === 'pending'));
const approval = (await controller.api('/sessions/' + session.id)).approvals.find(value => value.status === 'pending');
const approvalRejection = await setup(secondSetup, 'approval-block', 1); assert.equal((await controller.api('/sessions/' + session.id)).approvals.find(value => value.status === 'pending').id, approval.id);
await controller.api('/sessions/' + session.id + '/stop', { method: 'POST', input: {} }); record(stage, { ...approvalRejection, approvalId: approval.id });
stage = 'actual manifest fault installer preserves original pointer, PID and all data'; const snapshot = snapshotPersistentData(dataRoot), pointer = fs.readFileSync(path.join(installRoot, 'current.json'));
const faultResult = await setup(faultSetup, 'candidate-fault', 1);
assert.deepEqual(fs.readFileSync(path.join(installRoot, 'current.json')), pointer); assert.deepEqual(snapshotPersistentData(dataRoot), snapshot); assert.equal((await controller.connect()).identity.pid, original.pid); assert.ok(fs.existsSync(path.join(installRoot, 'AgentWorkbench.exe'))); record(stage, faultResult);
stage = 'actual pointer-published fault rolls original host and pointer back';
const originalBootstrap = fs.readFileSync(path.join(installRoot, 'AgentWorkbench.exe'));
const faultCandidate = path.join(installRoot, 'versions', 'owned-pointer-fault'); fs.cpSync(b, faultCandidate, { recursive: true });
await assert.rejects(ownedManager(faultCandidate).activate({ afterPointer: () => { throw Error('owned-post-pointer-fault'); } }), /owned-post-pointer-fault/);
assert.deepEqual(fs.readFileSync(path.join(installRoot, 'current.json')), pointer); assert.deepEqual(snapshotPersistentData(dataRoot), snapshot); assert.equal((await controller.connect()).identity.pid, original.pid); record(stage, { fixture: 'Explicit in-process afterPointer error with real retained host; original pointer bytes, database rows and artifact bytes checked.' });
assert.deepEqual(fs.readFileSync(path.join(installRoot, 'AgentWorkbench.exe')), originalBootstrap);
stage = 'actual update retains original PID, instance, DPAPI ciphertext and every persistent byte';
const updated = await setup(secondSetup, 'update'); const activeB = path.join(installRoot, 'versions', path.basename(b));
controller = new HostController({ appRoot: activeB, dataRoot, port, nodePath: process.execPath }); const after = (await controller.connect()).identity;
assert.equal(after.pid, original.pid); assert.equal(after.instanceId, original.instanceId); assert.deepEqual(snapshotPersistentData(dataRoot), snapshot); record(stage, { ...updated, hostPid: after.pid, instanceId: after.instanceId, data: snapshot });
assert.notDeepEqual(fs.readFileSync(path.join(installRoot, 'AgentWorkbench.exe')), originalBootstrap);
assert.deepEqual(fs.readFileSync(path.join(installRoot, 'AgentWorkbench.exe')), fs.readFileSync(path.join(activeB, 'AgentWorkbench.exe')));
stage = 'installed native bootstrap verifies and opens the registered release';
const nativeLauncher = spawn(path.join(installRoot, 'AgentWorkbench.exe'), ['--hidden-window'], { windowsHide: true, stdio: 'ignore', env });
await new Promise((resolve, reject) => { nativeLauncher.once('error', reject); nativeLauncher.once('exit', code => code === 0 ? resolve() : reject(Error('Owned native bootstrap failed (' + code + ')'))); });
await until(async () => (await controller.status()).desktopWindows.count === 1);
const nativeWindow = (await controller.status()).desktopWindows.entries[0], nativeProcess = observeWindowsProcess(nativeWindow.pid); assert.ok(nativeProcess); assert.equal((await controller.connect()).identity.pid, original.pid);
// A controlled WM_CLOSE to the owned main window is delegated to Electron by
// a separate automation connection below; no unknown process is terminated.
const child = spawn(path.join(activeB, 'desktop/AgentWorkbench.exe'), ['--app-root', activeB, '--data-root', dataRoot, '--port', String(port), '--hidden-window'], { cwd: activeB, env, windowsHide: true, stdio: 'ignore' });
await new Promise(resolve => child.once('exit', resolve)); assert.equal((await controller.status()).desktopWindows.count, 1);
record(stage, { electronPid: nativeWindow.pid, hostPid: original.pid, repeatLaunchReusedWindow: true });
// The native process has no debugger. Its owned renderer closes itself through
// the product's browser window action using a localhost, attested stop below.
stage = 'native desktop remains registered before explicit owned cleanup';
// Closing only this exactly observed test Electron tree after the launch check
// models a desktop crash. Host/persistent results remain; stale registration is
// removed through the normal identity check, without signalling a reused PID.
const observed = observeWindowsProcess(nativeWindow.pid); assert.equal(observed.createdAt, nativeWindow.createdAt);
await run('taskkill.exe', ['/PID', String(observed.pid), '/T', '/F'], { windowsHide: true, timeout: 15000, env });
await until(async () => (await controller.status()).desktopWindows.count === 0); assert.equal((await controller.connect()).identity.pid, original.pid);
record(stage, { fixture: 'Explicit owned Electron crash after exact actual PID/creation check; independent host retained.' });
stage = 'real host restart restores the owned saved credential and actual HTTP authentication';
const beforeRestart = snapshotPersistentData(dataRoot); await controller.stop(); const restarted = (await controller.start()).identity;
assert.notEqual(restarted.pid, original.pid); assert.deepEqual(snapshotPersistentData(dataRoot), beforeRestart);
await controller.api('/model-center/probe', { method: 'POST', input: { id: connectionId } }); assert.equal(authenticated, 3); record(stage, { originalPid: original.pid, restartedPid: restarted.pid, authenticatedRequests: authenticated });
if (['1.24.0', '1.24.1', '1.24.2', '1.24.3', '1.24.4', '1.24.5', '1.24.6', '1.24.7'].includes(version)) {
stage = 'actual installed current results and saved HTML user paths';
const pointer = JSON.parse(fs.readFileSync(path.join(installRoot, 'current.json')));
const installedRoot = fs.realpathSync.native(path.join(installRoot, 'versions', pointer.active.directory)), proofs = [];
for (const script of ['verify-file-results-ui.mjs', 'verify-local-files.mjs', 'verify-reading-reliability-ui.mjs']) {
const output = await run(process.execPath, [path.join(appRoot, 'scripts', script)], { cwd: appRoot,
env: { ...env, AGENT_WORKBENCH_QA_APP_ROOT: installedRoot }, windowsHide: true, timeout: 180000, maxBuffer: 1024 * 1024 });
const log = path.join(root, 'installed-' + script + '.log'); fs.writeFileSync(log, output.stdout + output.stderr);
const summary = JSON.parse(output.stdout.trim().split(/\r?\n/).findLast(line => line.startsWith('{"success":')));
const proofRoot = summary.directory ?? summary.root; assert.ok(proofRoot && path.resolve(proofRoot).startsWith(path.join(appRoot, '.qa') + path.sep));
const proof = JSON.parse(fs.readFileSync(path.join(proofRoot, 'result.json'), 'utf8'));
assert.equal(proof.success, true); if (script !== 'verify-reading-reliability-ui.mjs') assert.equal(proof.version, version); assert.equal(fs.realpathSync.native(proof.appRoot), installedRoot);
assert.equal(proof.checks.length, script === 'verify-file-results-ui.mjs' ? 16 : script === 'verify-local-files.mjs' ? 9 : 10); assert.deepEqual(proof.errors, []);
assert.equal(proof.realModelRequests, 0); proofs.push({ script, proof, log, sha256: hash(fs.readFileSync(log)) });
}
installedUI = { success: true, version, installedRoot, packagedSource, proofs, realModelRequests: 0 };
fs.writeFileSync('.qa/existing-results-installed-ui.json', JSON.stringify(installedUI, null, 2)); record(stage, installedUI);
}
stage = 'actual uninstall closes only the idle owned host and preserves user data';
const uninstallSnapshot = snapshotPersistentData(dataRoot), uninstallLog = path.join(root, 'uninstall.log');
await run(path.join(installRoot, 'unins000.exe'), ['/VERYSILENT', '/SUPPRESSMSGBOXES', '/NORESTART', '/LOG=' + uninstallLog], { windowsHide: true, timeout: 180000, env });
assert.equal(observeWindowsProcess(restarted.pid), null); assert.deepEqual(snapshotPersistentData(dataRoot), uninstallSnapshot); assert.equal(fs.existsSync(path.join(installRoot, 'AgentWorkbench.exe')), false); assert.equal(fs.existsSync(path.join(installRoot, 'current.json')), false);
record(stage, { log: uninstallLog, originalDataRoot: dataRoot, data: uninstallSnapshot });
report = { success: true, version, root, installRoot, dataRoot, port, appId, packagedSource, results, installedUI, realModelRequests: 0, scope: 'Actual isolated user-level Windows installers and shipped Electron/Node runtimes; own credentials and HTTP fixture only. QA appId, corruption, pointer-fault and crash fixtures explicitly recorded.' };
} catch (error) { report = { success: false, version, root, installRoot, dataRoot, port, stage, error: error.stack, results, realModelRequests: 0 }; throw error; }
finally {
releaseRequest?.(); await application?.close().catch(() => {});
if (controller) { try { const status = await controller.status(); if (!status.desktopWindows.count) await controller.stop(); } catch {} }
if (ownedModel) { ownedModel.closeAllConnections(); await new Promise(resolve => ownedModel.close(resolve)); }
fs.writeFileSync(path.join(root, 'result.json'), JSON.stringify(report, null, 2)); fs.writeFileSync('.qa/desktop-install-location.json', JSON.stringify({ root, result: path.join(root, 'result.json') })); console.log('EVIDENCE ' + root);
}
@@ -0,0 +1,15 @@
import fs from 'node:fs';
import assert from 'node:assert/strict';
const root='.qa/reading-reliability-s/native',origin='http://127.0.0.1:4318';
const bootstrap=await(await fetch(origin+'/api/bootstrap')).json();
const health=await(await fetch(origin+'/api/health')).json();
const status=await(await fetch(origin+'/api/host/status',{headers:{'X-Workbench-Token':bootstrap.token}})).json();
assert.equal(health.version,'1.24.7');assert.equal(health.pid,36552);assert.equal(status.desktopWindows.count,1);assert.equal(status.activity.active,false);
const initial=fs.readFileSync(root+'/after.txt','utf8'),center=fs.readFileSync(root+'/model-center.txt','utf8');
assert.match(initial,/release=1\.24\.7/);assert.match(initial,/v1\.24\.7/);assert.match(center,/高性能/);
for(const file of['after.jpg','model-center.jpg'])assert.equal(fs.readFileSync(root+'/'+file).subarray(0,3).toString('hex'),'ffd8ff');
const result={success:true,version:health.version,pid:health.pid,desktopWindows:status.desktopWindows.count,runtimeRoot:bootstrap.runtimeRoot,
observed:['Actual formal installed launcher opens unique 1.24.7 native window','Connected UI and original high-performance pool visible','No model dispatch, settings save, task start or original history mutation'],
screenshots:['after.jpg','model-center.jpg'],realModelRequests:0,
scope:'Computer Use observation from the actual installed original entry, not Electron fixture. Original R budget behavior is separately verified by write-blocked production web UI. State reconciliation follows after native review.'};
fs.writeFileSync(root+'/result.json',JSON.stringify(result,null,2)+'\n',{flag:'wx'});console.log(JSON.stringify(result));
@@ -0,0 +1,94 @@
import fs from 'node:fs';
import path from 'node:path';
import assert from 'node:assert/strict';
import { createHash } from 'node:crypto';
import { execFileSync } from 'node:child_process';
if(process.argv.includes('--help')){console.log('node scripts/verify-reading-s-evidence.mjs [--index] — read-only byte and result reconciliation for S. No network, model, installer or original driver replay.');process.exit(0);}
assert.ok(process.argv.length===2 || process.argv.length===3 && process.argv[2]==='--index');
const index=process.argv.includes('--index'),root=path.resolve('docs/reading-s-evidence');
const hash=bytes=>createHash('sha256').update(bytes).digest('hex');
const read=relative=>JSON.parse(fs.readFileSync(path.join(root,relative),'utf8').replace(/^\uFEFF/,''));
const manifest=read('manifest.json'),names=new Set();let total=0,indexFiles=0,indexBytes=0;
for(const entry of manifest.entries){
assert.equal(names.has(entry.path),false);names.add(entry.path);
const file=path.resolve(root,entry.path);assert.ok(file.startsWith(root+path.sep));
const bytes=fs.readFileSync(file);assert.equal(bytes.length,entry.bytes);assert.equal(hash(bytes),entry.sha256);total+=bytes.length;
if(entry.path.endsWith('.png'))assert.equal(bytes.subarray(0,8).toString('hex'),'89504e470d0a1a0a',entry.path);
if(/\.jpe?g$/.test(entry.path))assert.equal(bytes.subarray(0,3).toString('hex'),'ffd8ff',entry.path);
}
const walk=directory=>fs.readdirSync(directory,{withFileTypes:true}).flatMap(entry=>{
const file=path.join(directory,entry.name);assert.equal(fs.lstatSync(file).isSymbolicLink(),false,'No evidence links');
return entry.isDirectory()?walk(file):[path.relative(root,file).replaceAll('\\','/')];
});
assert.deepEqual(walk(root).filter(name=>name!=='manifest.json').sort(),[...names].sort());
assert.equal(manifest.count,names.size);assert.equal(manifest.bytes,total);assert.equal(manifest.realModelRequests,0);
const freeze=read('receipts/release/product-freeze.json'),identity=read('release-identity.json');
assert.equal(freeze.sourceCommit,manifest.sourceCommit);assert.equal(identity.sourceCommit,manifest.sourceCommit);assert.equal(identity.version,'1.24.7');
assert.equal(freeze.files.length,560);assert.equal(freeze.compiled.length,22);
for(const item of freeze.compiled)assert.equal(hash(fs.readFileSync(path.join(root,'compiled',item.path))),item.sha256);
for(const item of freeze.files.filter(item=>names.has('source-bytes/'+item.path)))assert.equal(hash(fs.readFileSync(path.join(root,'source-bytes',item.path))),item.sha256);
const build=read('validation/rebuild/result.json'),extra=read('receipts/release/extra-example-inputs.json');
assert.equal(build.success,true);assert.equal(build.sourceCommit,manifest.sourceCommit);assert.equal(build.inputCount,565);
assert.equal(build.compiled.length,22);assert.deepEqual(build.mismatches,[]);assert.equal(extra.success,true);assert.equal(extra.extra.length,5);
for(const file of freeze.files)assert.equal(build.inputs.find(item=>item.path===file.path)?.sha256,file.sha256,file.path);
for(const item of build.compiled)assert.deepEqual(item.actual,item.expected);
for(const name of ['archives','portable'])assert.equal(read('validation/'+name+'.json').success,true);
assert.equal(read('receipts/release/portable.json').success,false,'Keep initial portable startup rejection');
assert.equal(read('receipts/release/portable-first-failure-review.json').success,true);
assert.equal(read('receipts/release/portable-recheck.json').success,true);
for(const name of ['source-ui','portable-ui','installed-ui']){
const proof=read('validation/'+name+'/result.json');assert.equal(proof.success,true);assert.equal(proof.checks.length,10);
assert.equal(proof.realModelRequests,0);assert.deepEqual(proof.errors,[]);
}
const installation=read('validation/install/result.json');assert.equal(installation.success,true);assert.equal(installation.version,'1.24.7');
assert.equal(installation.results.length,12);assert.equal(installation.realModelRequests,0);assert.ok(installation.results.every(item=>item.passed));
assert.equal(installation.installedUI.proofs.length,3);
for(const [name,count]of[['installed-results',16],['installed-files',9]]){
const proof=read('validation/'+name+'/result.json');assert.equal(proof.success,true);assert.equal(proof.checks.length,count);assert.deepEqual(proof.errors,[]);
}
for(const [file,count]of[['final-core.log',96],['upgrade-regressions.log',18]]){
const log=fs.readFileSync(path.join(root,'receipts',file),'utf8');assert.match(log,new RegExp('pass '+count+'\\b'));assert.match(log,/fail 0\b/);
}
const provenance=read('receipts/replay-provenance.json');assert.equal(provenance.success,true);assert.equal(provenance.oldInputs.length,225);
for(const item of provenance.oldInputs)assert.equal(hash(fs.readFileSync(path.join(root,'source-before',item.path))),item.sha256);
for(const item of provenance.rRefs){
const original=fs.readFileSync(path.join('docs/live-delivery-r-evidence',item.path));assert.equal(hash(original),item.sha256);
assert.equal(hash(fs.readFileSync(path.join(root,'original-r',item.path))),item.sha256);
}
assert.equal(hash(fs.readFileSync(provenance.originalRManifest.path)),provenance.originalRManifest.sha256);
const beforeReplay=read('receipts/replay-before-final.json'),afterReplay=read('receipts/replay-after-final.json');
assert.equal(beforeReplay.externalModelRequests,0);assert.equal(afterReplay.externalModelRequests,0);
assert.equal(beforeReplay.fixtureRequests,8);assert.equal(afterReplay.fixtureRequests,6);
assert.equal(beforeReplay.stopReason,'step_limit');assert.equal(afterReplay.stopReason,'verification_incomplete');
assert.equal(beforeReplay.originalSourceSha,'');assert.equal(afterReplay.originalSourceSha,'');
assert.equal(afterReplay.goal,beforeReplay.goal);assert.deepEqual(afterReplay.budget,beforeReplay.budget);
const source=fs.readFileSync(path.join(root,'original-r/run/node/source-before/test/expense.test.mjs'),'utf8');
assert.equal(afterReplay.pages.map(item=>item.content).join(''),source);assert.equal(source.length,2584);
assert.equal(beforeReplay.pages.at(-1).next_offset,2173);assert.equal(afterReplay.pages.at(-1).next_offset,2584);
const initial=read('receipts/production-before.json'),before=read('receipts/formal-release/private-before.json');
function sameOriginal(a,b){
for(const key of['pid','hasKey','configSha256','projectsSha256','sessionsSha256','eventCursor','exports','persistent','modelCenterSha256'])assert.deepEqual(b[key],a[key],key);
assert.deepEqual(b.database.tables,a.database.tables);assert.deepEqual(b.database.sessionIds,a.database.sessionIds);
}
sameOriginal(initial,before);
for(const name of['install','final']){
const result=read('receipts/formal-release/'+name+'.json');assert.equal(result.success,true);assert.equal(result.pid,36552);assert.equal(result.version,'1.24.7');assert.equal(result.inspectorClosed,true);
sameOriginal(before,read('receipts/formal-release/'+name+'-private-after.json'));
}
assert.equal(read('receipts/release/browser-review/result.json').success,false,'Retain initial stale composer selector failure');
assert.equal(read('receipts/release/browser-review-final/result.json').success,false,'Retain narrow hidden-sidebar visibility wait failure');
assert.equal(read('receipts/release/browser-review-complete/result.json').success,false,'Retain initial model-center loading race');
const browser=read('receipts/release/browser-review-verified/result.json');assert.equal(browser.success,true);assert.equal(browser.views.length,2);
assert.deepEqual(browser.errors,[]);assert.deepEqual(browser.rejectedWrites,[]);
for(const view of browser.views){assert.equal(view.input,12352);assert.equal(view.reservation,29623);assert.equal(view.remaining,17429);}
const native=read('receipts/native/result.json');assert.equal(native.success,true);assert.equal(native.version,'1.24.7');assert.equal(native.realModelRequests,0);
if(index)for(const relative of [...names,'manifest.json']){
const gitPath='docs/reading-s-evidence/'+relative,bytes=fs.readFileSync(path.join(root,relative));
const blob=execFileSync('git',['cat-file','blob',':'+gitPath],{maxBuffer:Math.max(1024*1024,bytes.length+1024)});
assert.deepEqual(blob,bytes,gitPath);indexFiles++;indexBytes+=blob.length;
}
console.log(JSON.stringify({success:true,version:identity.version,sourceCommit:manifest.sourceCommit,
manifestSha256:hash(fs.readFileSync(path.join(root,'manifest.json'))),count:names.size,bytes:total,inputs:build.inputCount,compiled:22,
formalPid:36552,originalProjects:initial.projects,originalSessions:initial.database.sessionIds.length,
originalRowsExportsArtifactsAndModelCenter:'exact',offlineReplay:{before:8,after:6,realDeliveryPassed:false},realModelRequests:0,
...(index?{indexFiles,indexBytes}:{}),scope:'Read-only sealed byte and semantic reconciliation; no original driver or configured model replay.'}));
@@ -0,0 +1,13 @@
import fs from 'node:fs';
import assert from 'node:assert/strict';
import { execFileSync } from 'node:child_process';
const directory = '.qa/release-2066d209-2aa2-49ab-8f36-5c7d3987b8e1';
const entries = fs.readdirSync(directory); assert.deepEqual(entries, []);
const raw = execFileSync('powershell.exe', ['-NoProfile','-Command', "@(Get-CimInstance Win32_Process | Where-Object { $_.Name -eq 'node.exe' -and $_.CommandLine -like '*release-2066d209-2aa2-49ab-8f36-5c7d3987b8e1*' } | Select-Object ProcessId,Name) | ConvertTo-Json -Compress"], {encoding:'utf8',windowsHide:true}).trim();
assert.ok(raw === '' || raw === '[]');
const result = { success: true, firstPortableSucceeded: false, directory, entries, ownedHostProcesses: 0,
outcomeKnown: 'Initial health observation rejected before start: no host logs, launch intent, service-info or registration were created, and no Node command references the owned data root.',
cause: 'HOST_PORT_UNVERIFIED; underlying socket timing or port collision not retained by original CLI and is not proven. Candidate and protections unchanged.',
next: 'One separately recorded isolated serial portable verification with new owned UUID; never replay a production installation or configured model request.', realModelRequests:0 };
fs.writeFileSync('.qa/reading-reliability-s/release/portable-first-failure-review.json',JSON.stringify(result,null,2)+'\n',{flag:'wx'});
console.log(JSON.stringify(result));
@@ -0,0 +1,23 @@
import fs from 'node:fs';
import assert from 'node:assert/strict';
import { createHash } from 'node:crypto';
const original = fs.readFileSync('scripts/verify-desktop-install.mjs', 'utf8');
const edits = [
["from './bundle-desktop.mjs'", "from '../scripts/bundle-desktop.mjs'"],
["from './build-installer.mjs'", "from '../scripts/build-installer.mjs'"],
["['verify-file-results-ui.mjs', 'verify-local-files.mjs']", "['verify-file-results-ui.mjs', 'verify-local-files.mjs', 'verify-reading-reliability-ui.mjs']"],
['assert.equal(proof.version, version);', "if (script !== 'verify-reading-reliability-ui.mjs') assert.equal(proof.version, version);"],
["script === 'verify-file-results-ui.mjs' ? 16 : 9", "script === 'verify-file-results-ui.mjs' ? 16 : script === 'verify-local-files.mjs' ? 9 : 10"]
];
let body = original;
for (const [before, after] of edits) {
assert.equal(body.split(before).length, 2, before);
body = body.replace(before, after);
}
fs.writeFileSync('.qa/reading-s-install-lifecycle.mjs', body, { flag: 'wx' });
fs.writeFileSync('.qa/reading-reliability-s/release/install-driver-adaptation.json', JSON.stringify({
original: 'scripts/verify-desktop-install.mjs', sha256: createHash('sha256').update(original).digest('hex'),
target: '.qa/reading-s-install-lifecycle.mjs', edits,
scope: 'Identical actual install lifecycle with relative import fixes and extra ten-case S UI proof before uninstall. No product or frozen input changes.', realModelRequests: 0
}, null, 2) + '\n', { flag: 'wx' });
console.log('Prepared isolated installation driver.');
@@ -0,0 +1,7 @@
import fs from 'node:fs';
const r=JSON.parse(fs.readFileSync('.qa/reading-reliability-s/formal-release/install.json'));
const stages=r.operation.history.filter(x=>x.elapsedMs>=1000).map(({stage,scope,elapsedMs})=>({stage,scope,elapsedMs}));
fs.writeFileSync('.qa/reading-reliability-s/release/update-timing.json',JSON.stringify({version:r.version,
operationMs:r.operation.elapsedMs,totalDriverMs:Date.parse(r.finishedAt)-Date.parse(r.startedAt),stages,
scope:'Observed S actual formal update; full manifest/directory checks retained. Optimization is a separate planned slice.'},null,2)+'\n',{flag:'wx'});
console.log(JSON.stringify({version:r.version,operationMs:r.operation.elapsedMs,stages:stages.length}));
@@ -0,0 +1,48 @@
import fs from 'node:fs';
import path from 'node:path';
import assert from 'node:assert/strict';
import { createHash } from 'node:crypto';
const read = file => JSON.parse(fs.readFileSync(file, 'utf8').replace(/^\uFEFF/, ''));
const hash = file => createHash('sha256').update(fs.readFileSync(file)).digest('hex');
const root = '.qa/reading-reliability-s', freeze = read(root + '/release/product-freeze.json');
const release = read(root + '/release/candidate.json'); assert.deepEqual(read('releases/latest.json'), release);
assert.equal(release.documentation.sourceCommit, freeze.sourceCommit);
for (const item of freeze.compiled) assert.equal(hash(path.join(release.portableRoot, item.path)), item.sha256);
const roots = { 'source-ui': path.resolve('.qa/reading-s-ui-27708227-5a32-4705-b87a-d03aa2a753c8'),
'source-ui-first': path.resolve('.qa/reading-s-ui-476c68bd-100c-40e8-bd74-d47927e7b77f'),
'source-ui-corrected': path.resolve('.qa/reading-s-ui-a51eca8b-b2c5-4ac1-813e-4208927b39bf'),
'history-context': path.resolve('.qa/context-ui-008294dc-c99b-4373-8c4a-0c45eda7bee6'),
'history-budget': path.resolve('.qa/model-budget-ui-25db6351-c80f-4abb-9286-d10b0c3c1ff0'),
'history-selection': path.resolve('.qa/model-selection-k-d11c8059-7775-4e7b-b168-e7ba5a123ae1') };
for (const name of ['rebuild','portable-ui','install','archives','portable-recheck']) {
const receipt = read(root + '/release/' + name + '.json'); assert.equal(receipt.success, true);
assert.equal(receipt.sourceCommit, freeze.sourceCommit); assert.equal(receipt.candidate, release.portableRoot);
if (['rebuild','portable-ui','install'].includes(name)) roots[name] = path.dirname(receipt.resultPath);
}
const installation = read(path.join(roots.install, 'result.json')); assert.equal(installation.results.length, 12);
assert.ok(installation.results.every(item => item.passed));
for (const proof of installation.installedUI.proofs) {
const summary = fs.readFileSync(proof.log, 'utf8').trim().split(/\r?\n/).findLast(line => line.startsWith('{"success":'));
const parsed = JSON.parse(summary), value = parsed.directory ?? parsed.root;
const key = proof.script === 'verify-file-results-ui.mjs' ? 'installed-results' : proof.script === 'verify-local-files.mjs' ? 'installed-files' : 'installed-ui';
roots[key] = value;
}
for (const name of ['source-ui','portable-ui','installed-ui']) {
const proof = read(path.join(roots[name], 'result.json')); assert.equal(proof.success, true);
assert.equal(proof.checks.length, 10); assert.equal(proof.realModelRequests, 0); assert.deepEqual(proof.errors, []);
}
const build = read(path.join(roots.rebuild, 'result.json')); assert.equal(build.success, true); assert.equal(build.compiled.length, 22);
assert.equal(build.inputCount, freeze.files.length + 5);
for (const file of freeze.files) assert.equal(build.inputs.find(item => item.path === file.path)?.sha256, file.sha256, file.path);
assert.equal(read(root + '/release/extra-example-inputs.json').success, true);
for (const [file,count] of [['final-core.log',96],['upgrade-regressions.log',18]]) {
const log=fs.readFileSync(root+'/'+file,'utf8'); assert.match(log,new RegExp('pass '+count+'\\b')); assert.match(log,/fail 0\b/);
}
const output = { success: true, version: '1.24.7', sourceCommit: freeze.sourceCommit, candidate: release.portableRoot,
trackedInputs: freeze.files.length, independentInputs: build.inputCount, compiled: build.compiled.length,
isolatedInstallGroups: 12, sourcePortableInstalledReadingUI: '10+10+10', actualInstalledFilesAndResults: '16+9',
realModelRequests: 0, roots,
scope: 'All candidate gates bound to exact immutable source and portable; original Provider/data retention still requires actual formal upgrade.' };
fs.writeFileSync(root + '/validation-roots.json', JSON.stringify(roots,null,2)+'\n',{flag:'wx'});
fs.writeFileSync(root + '/release/preinstall-review.json', JSON.stringify(output,null,2)+'\n',{flag:'wx'});
console.log(JSON.stringify(output));
@@ -0,0 +1,43 @@
import fs from 'node:fs';
import path from 'node:path';
import assert from 'node:assert/strict';
import { execFileSync } from 'node:child_process';
import { createHash } from 'node:crypto';
const read = file=>JSON.parse(fs.readFileSync(file,'utf8'));
const hash = bytes=>createHash('sha256').update(bytes).digest('hex');
const root='.qa/reading-reliability-s', oldRoot=root+'/source-before';
const oldCommit=execFileSync('git',['rev-parse','fbedb738'],{encoding:'utf8'}).trim();
const paths=execFileSync('git',['ls-tree','-r','--name-only',oldCommit,'--','src','package.json'],{encoding:'utf8'}).trim().split('\n');
const oldInputs=[];
for(const relative of paths){
const original=execFileSync('git',['show',oldCommit+':'+relative],{maxBuffer:10*1024*1024});
const actual=fs.readFileSync(path.join(oldRoot,relative)); assert.deepEqual(actual,original,relative);
oldInputs.push({path:relative,bytes:actual.length,sha256:hash(actual)});
}
assert.equal(fs.lstatSync(oldRoot+'/node_modules').isSymbolicLink(),true);
const rManifestPath='docs/live-delivery-r-evidence/manifest.json', rManifest=read(rManifestPath);
const rRefs=[];
for(const entry of rManifest.entries.filter(entry=>entry.source.replaceAll('\\','/').includes('/live-delivery-r/node/') &&
(entry.source.replaceAll('\\','/').endsWith('/node/export.json') || entry.source.replaceAll('\\','/').includes('/node/source-before/')))){
const bytes=fs.readFileSync(entry.source); assert.equal(bytes.length,entry.bytes); assert.equal(hash(bytes),entry.sha256);
assert.equal(hash(fs.readFileSync(path.join('docs/live-delivery-r-evidence',entry.path))),entry.sha256);
rRefs.push({path:entry.path,source:entry.source,bytes:entry.bytes,sha256:entry.sha256});
}
assert.ok(rRefs.some(entry=>entry.source.replaceAll('\\','/').endsWith('/node/export.json')));
const original=read('.qa/live-delivery-r/node/export.json');
const before=read(root+'/replay-before-final.json'),after=read(root+'/replay-after-final.json');
assert.equal(before.goal,original.execution.task.goal); assert.equal(after.goal,before.goal);assert.deepEqual(after.budget,before.budget);
assert.equal(before.originalSourceSha,'');assert.equal(after.originalSourceSha,'');
const source=fs.readFileSync('.qa/live-delivery-r/node/source-before/test/expense.test.mjs','utf8');
assert.equal(after.pages.map(page=>page.content).join(''),source);
assert.equal(after.pages.at(-1).next_offset,source.length);
assert.ok(before.pages.at(-1).next_offset<source.length);
const result={success:true,oldCommit,oldInputs,sourceBeforeNodeModules:'junction intentionally excluded; no recursive traversal',
originalRManifest:{path:rManifestPath,sha256:hash(fs.readFileSync(rManifestPath))},rRefs,
source:{characters:source.length,sha256:hash(Buffer.from(source))},
before:{requests:before.fixtureRequests,stopReason:before.stopReason,lastVisibleOffset:before.pages.at(-1).next_offset},
after:{requests:after.fixtureRequests,stopReason:after.stopReason,lastVisibleOffset:after.pages.at(-1).next_offset,fullSourceMatched:true},
originalAssistantReplies:original.session.providerMessages.filter(message=>message.role==='assistant').slice(0,3).map(message=>({sha256:hash(JSON.stringify(message))})),
sourceChanges:0,externalModelRequests:0,realModelDeliveryPassed:false,purgeBenefitMeasured:false};
fs.writeFileSync(root+'/replay-provenance.json',JSON.stringify(result,null,2)+'\n',{flag:'wx'});
console.log(JSON.stringify({success:true,oldInputs:oldInputs.length,rRefs:rRefs.length,before:result.before,after:result.after}));
@@ -0,0 +1,44 @@
import fs from 'node:fs';
import path from 'node:path';
import assert from 'node:assert/strict';
import { execFileSync } from 'node:child_process';
import { createHash } from 'node:crypto';
const read = file=>JSON.parse(fs.readFileSync(file,'utf8'));
const hash = bytes=>createHash('sha256').update(bytes).digest('hex');
const root='.qa/reading-reliability-s', oldRoot=root+'/source-before';
const oldCommit=execFileSync('git',['rev-parse','fbedb738'],{encoding:'utf8'}).trim();
const paths=execFileSync('git',['ls-tree','-r','--name-only',oldCommit,'--','src','package.json'],{encoding:'utf8'}).trim().split('\n');
const oldInputs=[];
for(const relative of paths){
const original=execFileSync('git',['show',oldCommit+':'+relative],{maxBuffer:10*1024*1024});
const actual=fs.readFileSync(path.join(oldRoot,relative)); assert.deepEqual(actual,original,relative);
oldInputs.push({path:relative,bytes:actual.length,sha256:hash(actual)});
}
assert.equal(fs.lstatSync(oldRoot+'/node_modules').isSymbolicLink(),true);
const rManifestPath='docs/live-delivery-r-evidence/manifest.json', rManifest=read(rManifestPath);
const rRefs=[];
for(const entry of rManifest.files.filter(entry=>entry.path.startsWith('run/node/') &&
(entry.path.endsWith('/node/export.json') || entry.path.includes('/node/source-before/')))){
const source=path.resolve('.qa/live-delivery-r',entry.path.slice('run/'.length));
const bytes=fs.readFileSync(source); assert.equal(bytes.length,entry.bytes); assert.equal(hash(bytes),entry.sha256);
assert.equal(hash(fs.readFileSync(path.join('docs/live-delivery-r-evidence',entry.path))),entry.sha256);
rRefs.push({path:entry.path,source,bytes:entry.bytes,sha256:entry.sha256});
}
assert.ok(rRefs.some(entry=>entry.source.replaceAll('\\','/').endsWith('/node/export.json')));
const original=read('.qa/live-delivery-r/node/export.json');
const before=read(root+'/replay-before-final.json'),after=read(root+'/replay-after-final.json');
assert.equal(before.goal,original.execution.task.goal); assert.equal(after.goal,before.goal);assert.deepEqual(after.budget,before.budget);
assert.equal(before.originalSourceSha,'');assert.equal(after.originalSourceSha,'');
const source=fs.readFileSync('.qa/live-delivery-r/node/source-before/test/expense.test.mjs','utf8');
assert.equal(after.pages.map(page=>page.content).join(''),source);
assert.equal(after.pages.at(-1).next_offset,source.length);
assert.ok(before.pages.at(-1).next_offset<source.length);
const result={success:true,oldCommit,oldInputs,sourceBeforeNodeModules:'junction intentionally excluded; no recursive traversal',
originalRManifest:{path:rManifestPath,sha256:hash(fs.readFileSync(rManifestPath))},rRefs,
source:{characters:source.length,sha256:hash(Buffer.from(source))},
before:{requests:before.fixtureRequests,stopReason:before.stopReason,lastVisibleOffset:before.pages.at(-1).next_offset},
after:{requests:after.fixtureRequests,stopReason:after.stopReason,lastVisibleOffset:after.pages.at(-1).next_offset,fullSourceMatched:true},
originalAssistantReplies:original.session.providerMessages.filter(message=>message.role==='assistant').slice(0,3).map(message=>({sha256:hash(JSON.stringify(message))})),
sourceChanges:0,externalModelRequests:0,realModelDeliveryPassed:false,purgeBenefitMeasured:false};
fs.writeFileSync(root+'/replay-provenance.json',JSON.stringify(result,null,2)+'\n',{flag:'wx'});
console.log(JSON.stringify({success:true,oldInputs:oldInputs.length,rRefs:rRefs.length,before:result.before,after:result.after}));
@@ -0,0 +1,84 @@
import fs from 'node:fs';
import path from 'node:path';
import assert from 'node:assert/strict';
import { execFileSync } from 'node:child_process';
import { pathToFileURL } from 'node:url';
import test from 'node:test';
import { fixture, compatibleFixture, completion, call, until, assertToolPairs } from '../tests/helpers.mjs';
const mode = process.argv[2]; assert.ok(['before', 'after'].includes(mode));
const root = path.resolve('.qa/reading-reliability-s');
const label = process.argv[3] ?? 'final'; assert.match(label, /^[a-z0-9-]+$/);
const output = path.join(root, 'replay-' + mode + '-' + label + '.json'); assert.ok(!fs.existsSync(output));
let sourceRoot = path.resolve('.');
if (mode === 'before') {
sourceRoot = path.join(root, 'source-before');
if (!fs.existsSync(sourceRoot)) { fs.mkdirSync(sourceRoot);
const names = execFileSync('git', ['ls-tree', '-r', '--name-only', 'fbedb738', '--', 'src', 'package.json'], { encoding: 'utf8' }).trim().split('\n');
for (const name of names) { const destination = path.join(sourceRoot, name); fs.mkdirSync(path.dirname(destination), { recursive: true });
fs.writeFileSync(destination, execFileSync('git', ['show', 'fbedb738:' + name], { maxBuffer: 10 * 1024 * 1024 })); }
fs.symlinkSync(path.resolve('node_modules'), path.join(sourceRoot, 'node_modules'), 'junction');
}
}
const load = name => import(pathToFileURL(path.join(sourceRoot, name)));
const { Runtime } = await load('src/server/runtime.mjs'), { Provider } = await load('src/server/provider.mjs');
const { verifyPlanEvidence } = await load('src/core/context/request-plan.mjs');
const original = JSON.parse(fs.readFileSync('.qa/live-delivery-r/node/export.json', 'utf8'));
const replies = original.session.providerMessages.filter(message => message.role === 'assistant').slice(0, 3);
test('R original first three turns, then deterministic source continuation within the same original limits: ' + mode, { timeout: 60000 }, async t => {
const f = fixture(t); fs.cpSync('.qa/live-delivery-r/node/source-before', f.projectRoot, { recursive: true });
const git = args => execFileSync('git', ['-C', f.projectRoot, ...args], { encoding: 'utf8', windowsHide: true });
git(['init', '--initial-branch=main']); git(['config', 'core.autocrlf', 'false']); git(['config', 'user.name', 'QA']); git(['config', 'user.email', 'qa@workbench.invalid']);
git(['add', '.']); git(['commit', '-m', 'Original R isolated input']);
const requests = [], pages = [], toolCounts = [];
const model = await compatibleFixture(t, (_route, body) => {
assertToolPairs(body.messages); requests.push(body);
const n = requests.length;
const page = f.session.providerMessages.filter(m => m.role === 'tool').map(m => JSON.parse(m.content))
.filter(value => value.path === 'test/expense.test.mjs' && typeof value.content === 'string').at(-1);
if (page) {
if (!pages.some(value => value.offset === page.offset)) pages.push(page);
const sent = body.messages.some(message => {
if (message.role === 'tool') { try { const value = JSON.parse(message.content); return value.path === page.path && value.offset === page.offset && value.content === page.content; } catch {} }
if (message.role !== 'system') return false;
const line = message.content.split('\n').find(line => line.startsWith('{"working_set"'));
return line && JSON.parse(line).data.entries.some(entry => entry.observation.path === page.path && entry.observation.offset === page.offset && entry.content === page.content);
});
if (n > 4) assert.ok(sent, 'The entire most recent pure-read source page must reach the next bounded request.');
}
if (n <= 3) {
const reply = structuredClone(replies[n - 1]);
for (const tool of reply.tool_calls) if (tool.function.name === 'run_checks') {
const plan = [...f.store.core.check_plan.values()].at(-1);
assert.ok(plan); tool.function.arguments = JSON.stringify({ plan_id: plan.id });
}
return { body: { ...completion(), choices: [{ message: reply, finish_reason: 'tool_calls' }] } };
}
if (page.next_offset < page.total_characters) return { body: completion([call('s-next-' + n, 'read_file', {
path: page.path, start_line: page.range_start_line, end_line: page.range_end_line, offset: page.next_offset, limit: 16000,
...(mode === 'after' ? { expected_sha256: page.sha256 } : {}) })]) };
return { body: completion([], 'Isolated source coverage complete; no product delivery claimed.') };
});
f.store.saveConfig({ mode: 'compatible', baseUrl: model.baseUrl, model: 'gpt-6.1-sol', contextLimit: 16000,
maxOutputTokens: 2048, maxSteps: 8, maxRunTokens: 90000, maxTaskTokens: 90000, maxRunDurationMs: 600000 });
const runtime = new Runtime(f.store, new Provider({ retryPolicy: { attempts: 1 } })); f.onCleanup(() => runtime.stopAll());
runtime.start(f.session.id, original.execution.task.goal);
await until(async () => {
for (const approval of f.session.approvals.filter(value => value.status === 'pending')) runtime.decide(f.session.id, approval.id, true);
toolCounts.push([...f.store.core.tool.values()].length);
return !runtime.isActive(f.session.id);
}, 50000);
const plans = [...f.store.core.context_plan.values()]; for (const plan of plans) verifyPlanEvidence(f.store, f.session, plan);
const run = f.store.core.run.get(f.session.runId);
const result = { mode, sourceCommit: mode === 'before' ? 'fbedb738' : 'working-source', externalModelRequests: 0,
fixtureRequests: requests.length, status: f.session.status, stopReason: run.stopReason, pages, plans, requests,
tools: [...f.store.core.tool.values()], goal: original.execution.task.goal, budget: run.budget,
originalSourceSha: execFileSync('git', ['-C', f.projectRoot, 'status', '--porcelain'], { encoding: 'utf8' }) };
fs.writeFileSync(output, JSON.stringify(result, null, 2) + '\n', { flag: 'wx' });
console.log(JSON.stringify({ mode, requests: requests.length, status: result.status, stopReason: result.stopReason,
pages: pages.map(p => [p.offset, p.next_offset, p.maximum_message_bytes]), failure: f.session.failure }));
assert.ok(pages.length); assert.equal(result.originalSourceSha, '');
if (mode === 'after') { assert.equal(pages.at(-1).next_offset, pages.at(-1).total_characters); assert.ok(requests.length < 8);
assert.equal(run.stopReason, 'verification_incomplete'); assert.equal(f.session.failure, null); }
});
@@ -0,0 +1,63 @@
import fs from 'node:fs';
import path from 'node:path';
import assert from 'node:assert/strict';
import { createHash } from 'node:crypto';
const hash=bytes=>createHash('sha256').update(bytes).digest('hex');
const read=file=>JSON.parse(fs.readFileSync(file,'utf8').replace(/^\uFEFF/,''));
const base=path.resolve('docs/reading-s-evidence');assert.equal(fs.existsSync(base),false);fs.mkdirSync(base);
const entries=[];
function copy(file,relative){
assert.equal(fs.lstatSync(file).isSymbolicLink(),false,'Never copy a junction');
const target=path.resolve(base,relative);assert.ok(target.startsWith(base+path.sep));
const bytes=fs.readFileSync(file);fs.mkdirSync(path.dirname(target),{recursive:true});fs.writeFileSync(target,bytes,{flag:'wx'});
entries.push({path:relative.replaceAll('\\','/'),bytes:bytes.length,sha256:hash(bytes),source:path.resolve(file)});
}
function walk(root,destination){
for(const entry of fs.readdirSync(root,{withFileTypes:true})){
const source=path.join(root,entry.name),relative=path.posix.join(destination,entry.name);
assert.equal(fs.lstatSync(source).isSymbolicLink(),false,'No recursive links');
if(entry.isDirectory())walk(source,relative);else copy(source,relative);
}
}
const root='.qa/reading-reliability-s',freeze=read(root+'/release/product-freeze.json'),release=read(root+'/release/candidate.json');
assert.equal(freeze.version,'1.24.7');assert.equal(release.documentation.sourceCommit,freeze.sourceCommit);
assert.equal(read(root+'/formal-release/final.json').success,true);assert.equal(read(root+'/native/result.json').success,true);
// Explicitly skip source-before at the root: it has a node_modules junction.
for(const entry of fs.readdirSync(root,{withFileTypes:true})){
if(entry.name==='source-before')continue;
const source=path.join(root,entry.name),relative=path.posix.join('receipts',entry.name);
if(entry.isDirectory())walk(source,relative);else copy(source,relative);
}
for(const name of fs.readdirSync('.qa').filter(name=>/^reading-s-.*\.mjs$/.test(name)))copy('.qa/'+name,'drivers/'+name+'.txt');
copy('.qa/desktop-f-state.mjs','drivers/desktop-f-state.mjs.txt');
copy('.qa/reading-s-review-notes.json','review-notes.json');
copy('.qa/reading-s-ui-validation.json','validation/history-context-summary.json');
const provenance=read(root+'/replay-provenance.json');
for(const item of provenance.oldInputs){const file=path.join(root,'source-before',item.path);assert.equal(hash(fs.readFileSync(file)),item.sha256);copy(file,'source-before/'+item.path);}
for(const item of provenance.rRefs){assert.equal(hash(fs.readFileSync(item.source)),item.sha256);copy(item.source,'original-r/'+item.path);}
for(const item of freeze.compiled){const file=path.join(release.portableRoot,item.path);assert.equal(hash(fs.readFileSync(file)),item.sha256);copy(file,'compiled/'+item.path);}
for(const item of freeze.files.filter(item=>/^src\//.test(item.path)||/reading-reliability|read-budget|working-set|budget-presentation|model-reservation|qa-evidence/.test(item.path))){
assert.equal(hash(fs.readFileSync(item.path)),item.sha256);copy(item.path,'source-bytes/'+item.path);
}
const roots=read(root+'/validation-roots.json');
for(const[name,directory]of Object.entries(roots)){
assert.ok(path.resolve(directory).startsWith(path.resolve('.qa')+path.sep));
for(const entry of fs.readdirSync(directory,{withFileTypes:true}))if(entry.isFile()&&/\.(?:json|png|jpg|log|txt)$/.test(entry.name))copy(path.join(directory,entry.name),'validation/'+name+'/'+entry.name);
}
// Preserve the exact context-history screenshots outside its random result root.
for(const name of fs.readdirSync('.qa/screenshots').filter(name=>/^reading-s-.*\.png$/.test(name)))copy('.qa/screenshots/'+name,'validation/history-context/screenshots/'+name);
copy(root+'/release/archives.json','validation/archive-gate.json');copy(root+'/release/portable-recheck.json','validation/portable-gate.json');
copy('.qa/archives-validation.json','validation/archives.json');copy('.qa/release-validation.json','validation/portable.json');
copy('releases/latest.json','release-latest.json');
copy(path.join(release.portableRoot,'MANIFEST.json'),'portable-manifest.json');copy(path.join(release.sourceRoot,'MANIFEST.json'),'source-manifest.json');
const identity={version:'1.24.7',sourceCommit:freeze.sourceCommit,candidate:path.basename(release.portableRoot),
portableManifestSha256:hash(fs.readFileSync(path.join(release.portableRoot,'MANIFEST.json'))),
sourceManifestSha256:hash(fs.readFileSync(path.join(release.sourceRoot,'MANIFEST.json'))),installer:release.installer,
archives:Object.fromEntries(['portable','source','evidence'].map(kind=>[kind,release[kind]])),realModelRequests:0};
fs.writeFileSync('docs/reading-s-release-identity.json',JSON.stringify(identity,null,2)+'\n',{flag:'wx'});
copy('docs/reading-s-release-identity.json','release-identity.json');
const manifest={schemaVersion:1,createdAt:new Date().toISOString(),sourceCommit:freeze.sourceCommit,version:'1.24.7',realModelRequests:0,
scope:'S source reading and reservation explanations, original failures, independent source/build/UI/installation and unchanged original production state. Historical R originals are referenced and matched. No credential values or external model calls. Frozen drivers are evidence, not replay instructions.',
count:entries.length,bytes:entries.reduce((sum,entry)=>sum+entry.bytes,0),entries:entries.sort((a,b)=>a.path.localeCompare(b.path))};
fs.writeFileSync(path.join(base,'manifest.json'),JSON.stringify(manifest,null,2)+'\n',{flag:'wx'});
console.log(JSON.stringify({success:true,count:manifest.count,bytes:manifest.bytes,manifestSha256:hash(fs.readFileSync(path.join(base,'manifest.json')))}));
@@ -0,0 +1,13 @@
import fs from 'node:fs';
import assert from 'node:assert/strict';
for (const file of ['package.json', 'package-lock.json']) {
const value = JSON.parse(fs.readFileSync(file)); assert.equal(value.version, '1.24.6'); value.version = '1.24.7';
if (value.packages?.['']) value.packages[''].version = '1.24.7';
fs.writeFileSync(file, JSON.stringify(value, null, 2) + '\n');
}
for (const file of ['scripts/upgrade-running.mjs', 'scripts/verify-all-ui.mjs', 'scripts/verify-controlled-release.mjs', 'scripts/verify-desktop-install.mjs', 'scripts/verify-release.mjs']) {
const old = fs.readFileSync(file, 'utf8'); assert.ok(old.includes("'1.24.6']"), file);
fs.writeFileSync(file, old.replaceAll("'1.24.6']", "'1.24.6', '1.24.7']"));
}
const file = 'src/server/diagnostics.mjs', old = fs.readFileSync(file, 'utf8');
assert.ok(old.includes("APP_VERSION = '1.24.6'")); fs.writeFileSync(file, old.replace("APP_VERSION = '1.24.6'", "APP_VERSION = '1.24.7'"));
File diff suppressed because it is too large Load Diff
File diff suppressed because one or more lines are too long
@@ -0,0 +1,3 @@
dist/
.awb-*
node_modules/
@@ -0,0 +1,2 @@
# Expense report acceptance
Repair the three source modules using the immutable original nine tests; do not edit checks or owner files.
@@ -0,0 +1 @@
Owner working copy differs from staging
@@ -0,0 +1 @@
Owner unstaged change
@@ -0,0 +1 @@
Owner untracked change
@@ -0,0 +1,9 @@
{
"name": "owned-live-expense-report",
"private": true,
"type": "module",
"scripts": {
"test": "node --test test/expense.test.mjs",
"build": "node scripts/build.mjs"
}
}
@@ -0,0 +1,8 @@
import { spawnSync } from 'node:child_process';
import { readFileSync } from 'node:fs';
for (const file of ['src/amount.mjs','src/summary.mjs','src/report.mjs']) {
const result=spawnSync(process.execPath,['--check',file],{encoding:'utf8'});
if(result.status!==0) { process.stderr.write(result.stderr);process.exit(result.status??1); }
if(!readFileSync(file,'utf8').trim()) throw new Error('Empty business module');
}
console.log('All three business modules compile.');
@@ -0,0 +1 @@
export function amountToCents(value) { return 0; }
@@ -0,0 +1 @@
export function formatReport(groups) { return ''; }
@@ -0,0 +1,2 @@
import { amountToCents } from './amount.mjs';
export function summarize(rows) { return []; }
@@ -0,0 +1,37 @@
import test from 'node:test';
import assert from 'node:assert/strict';
import { amountToCents } from '../src/amount.mjs';
import { summarize } from '../src/summary.mjs';
import { formatReport } from '../src/report.mjs';
test('decimal strings become exact cents including zero and refunds',()=>{
for(const [input,expected] of [['0',0],[' 12.34 ',1234],['+2.5',250],['-1.05',-105],['-0.00',0],['0009.10',910]]) assert.equal(amountToCents(input),expected);
});
test('finite decimal numbers use the same cents contract',()=>{
for(const [input,expected] of [[0,0],[1.01,101],[2.5,250],[-4.32,-432]]) assert.equal(amountToCents(input),expected);
});
test('malformed overprecision nonfinite and unsafe amounts are rejected',()=>{
for(const input of ['',null,undefined,true,{},'1e2','1,000','.50','1.234',1.234,NaN,Infinity,'900719925474099.99']) assert.throws(()=>amountToCents(input),TypeError,String(input));
});
test('independent currencies are sorted and summed without floating point loss',()=>{
assert.deepEqual(summarize([{currency:'usd',amount:'0.10'},{currency:' CNY ',amount:3},{currency:'USD',amount:'0.20'}]),[
{currency:'CNY',cents:300,count:1},{currency:'USD',cents:30,count:2}]);
});
test('refunds cancel totals while counting every original record',()=>{
assert.deepEqual(summarize([{currency:'eur',amount:'1.20'},{currency:'EUR',amount:'-1.20'}]),[{currency:'EUR',cents:0,count:2}]);
});
test('summarizing retains caller rows and handles an empty list',()=>{
const rows=Object.freeze([Object.freeze({currency:'usd',amount:'2.00'})]);
assert.deepEqual(summarize(rows),[{currency:'USD',cents:200,count:1}]);assert.deepEqual(summarize([]),[]);
});
test('invalid rows currency and unsafe aggregate totals are rejected',()=>{
for(const rows of [null,{},[null],[{currency:'US',amount:'1'}],[{currency:'U1D',amount:'1'}],[{currency:'USD',amount:'1.234'}],
[{currency:'USD',amount:'90071992547409.91'},{currency:'USD',amount:'0.01'}]]) assert.throws(()=>summarize(rows),TypeError);
});
test('reports use two decimals and render negative subunit amounts',()=>{
assert.equal(formatReport([{currency:'CNY',cents:300,count:2},{currency:'USD',cents:-105,count:1}]),'CNY 3.00 (2)\nUSD -1.05 (1)');
assert.equal(formatReport([{currency:'EUR',cents:-5,count:1}]),'EUR -0.05 (1)');assert.equal(formatReport([]),'');
});
test('complete expense flow preserves exact decimal sums',()=>{
assert.equal(formatReport(summarize([{currency:'usd',amount:'0.29'},{currency:'USD',amount:'-0.30'},{currency:'CNY',amount:'10.50'}])),
'CNY 10.50 (1)\nUSD -0.01 (2)');
});
File diff suppressed because it is too large Load Diff
@@ -0,0 +1,20 @@
> local-agent-workbench@1.24.7 build
> tsc --noEmit && vite build && node scripts/build-mobile.mjs && node scripts/build-server.mjs
vite v7.3.7 building client environment for production...
transforming...
✓ 1761 modules transformed.
rendering chunks...
computing gzip size...
dist-client/index.html 0.60 kB │ gzip: 0.41 kB
dist-client/.vite/manifest.json 0.65 kB │ gzip: 0.24 kB
dist-client/mobile.html 0.84 kB │ gzip: 0.47 kB
dist-client/assets/mobile-joLHctQw.css 12.39 kB │ gzip: 3.22 kB
dist-client/assets/desktop-DC8YSCxQ.css 173.13 kB │ gzip: 32.79 kB
dist-client/assets/mobile-DdtBN08G.js 26.97 kB │ gzip: 9.31 kB
dist-client/assets/approval-presentation-QyCkfhWb.js 227.03 kB │ gzip: 71.74 kB
dist-client/assets/desktop-CwtjTp8K.js 414.44 kB │ gzip: 117.80 kB
✓ built in 1.36s
Mobile shell built: 3 allowed assets; API and project data are never cached.
Server bundle built.
@@ -0,0 +1,20 @@
> local-agent-workbench@1.24.7 build
> tsc --noEmit && vite build && node scripts/build-mobile.mjs && node scripts/build-server.mjs
vite v7.3.7 building client environment for production...
transforming...
✓ 1761 modules transformed.
rendering chunks...
computing gzip size...
dist-client/index.html 0.60 kB │ gzip: 0.41 kB
dist-client/.vite/manifest.json 0.65 kB │ gzip: 0.24 kB
dist-client/mobile.html 0.84 kB │ gzip: 0.47 kB
dist-client/assets/mobile-joLHctQw.css 12.39 kB │ gzip: 3.22 kB
dist-client/assets/desktop-DC8YSCxQ.css 173.13 kB │ gzip: 32.79 kB
dist-client/assets/mobile-DdtBN08G.js 26.97 kB │ gzip: 9.31 kB
dist-client/assets/approval-presentation-QyCkfhWb.js 227.03 kB │ gzip: 71.74 kB
dist-client/assets/desktop-CwtjTp8K.js 414.44 kB │ gzip: 117.80 kB
✓ built in 1.28s
Mobile shell built: 3 allowed assets; API and project data are never cached.
Server bundle built.
@@ -0,0 +1,138 @@
✔ input that fits the engine window but exceeds remaining credit diagnoses the credit limit (0.5189ms)
✔ window overflow and historical pauses retain distinct and honest diagnoses (0.08ms)
✔ R observed reservation shortage keeps the window-fit diagnosis separate from Run/Task credit (0.0948ms)
✔ a budget pause persists concrete text/tool/history/image breakdown and never dispatches an oversized request (4374.2087ms)
✔ explicit frozen preview resumes the same Task in a new Run, preserves completed tools and does not change global config or selected model (4372.6778ms)
✔ insufficient budgets, silently clamped capacity and arbitrary config fields reject without a physical request or mutating task budget (2610.4391ms)
✔ changed session, changed registry, expiry and altered frozen preview all require a new explicit preview (2935.4509ms)
✔ preview preserves unknown invocation status and increasing Task allowance is committed only with confirmed new Run (2730.9842ms)
✔ reservation uses actual known usage from the same route, never lowers estimates or rewrites original receipts (1.7352ms)
✔ reservation verification binds original Task/Run consumption and rejects changed usage or foreign provenance (0.5443ms)
✔ known high input calibrates the next request and stops before dispatch while preserving the completed read (3316.1864ms)
✔ one reported overrun stops legacy requests and keeps actual use and excess (2669.9923ms)
✔ one reported overrun stops pool requests and keeps actual use and excess (2034.2209ms)
✔ one reported overrun stops connection requests and keeps actual use and excess (1604.1616ms)
✔ reported usage below the limit prevents an unaffordable legacy retry before another dispatch (1547.8615ms)
✔ reported usage below the limit prevents an unaffordable pool retry before another dispatch (2108.4577ms)
✔ native bridge seals calibrated refusal with original consumption and preserves versioned request verification (14538.4325ms)
✔ continuation reads advance through historical pages without a false pause while repeated current-Run pages still pause (1.8326ms)
✔ a new tool page replaces old latest-feedback space and still obeys its escaped protocol budget (0.4507ms)
✔ budgeted artifact pages count escaped protocol bytes and reconstruct original multilingual JSON without losing characters (9.4687ms)
✔ known artifact coverage detects semantic repeats despite changing limits, and advancing pages or direct evidence restore progress (0.4851ms)
✔ actual large artifact reads are present in the next bounded HTTP request and next_offset progresses without redoing tools (2748.0827ms)
✔ a real repeated artifact loop pauses the same Task after recorded results, survives reload and only resumes explicitly (3331.3846ms)
✔ actual HTTP continuation after reload sends distinct historically covered pages without false no_progress or replay (5011.4101ms)
✔ continuations distinguish unread source, later original line ranges and already returned result wrappers (501.4881ms)
✔ pure-file batch returns unused small-page space while mixed and historical allocation stays conservative (0.3313ms)
✖ chat-completions: complete actual Unicode/CRLF pages and small source survive the next bounded wire request (6433.4656ms)
✖ responses: complete actual Unicode/CRLF pages and small source survive the next bounded wire request (7626.023ms)
✖ anthropic-messages: complete actual Unicode/CRLF pages and small source survive the next bounded wire request (5372.4752ms)
✔ a real two-reply failure replay reaches request three at 16K with complete goal, original sources and sealed tool evidence (7310.6341ms)
✔ request deduplication retains long original intent, extra requirements, constraints and acceptance without changing the stored contract (1581.9827ms)
✔ only the routine context avoids plan echoes; diagnostic API and requests without an included plan keep original observations (578.4796ms)
✔ small code and actual failed command stay visible when their original multi-tool group cannot fit (1596.0836ms)
✔ mixed group retains six distinct small file observations with exact multilingual text and bounded request bytes (2027.4543ms)
✔ working observations are bound to the current Task and stop treating a pre-application file revision as current evidence (1752.102ms)
✔ a sealed observation rejects changed or missing original result artifacts instead of recreating evidence (1214.9148ms)
✔ a real passed-check observation is excluded after a source edit invalidates the check, while its historical sent evidence remains valid (5057.6264ms)
✔ budgeted file pages reconstruct UTF-8 text including long single lines and surrogate boundaries (6.0966ms)
✔ actual HTTP multi-tool reads survive an oversized runtime snapshot and the sealed evidence survives reopening (7643.6717ms)
✖ actual HTTP file pagination advances over a long file with each returned page visible in the next request (8186.3182ms)
✔ later mixed history pages cannot displace the actual module sources and complete requirements (4355.3822ms)
✔ a later narrow page of the same file revision cannot evict the wider actual observation (2552.4222ms)
✔ six small source files stay complete across metadata turns even when mandatory evidence exceeds the soft target (2409.7647ms)
✔ historical read-artifact navigation points to the original artifact and next page, never its result wrapper (966.8864ms)
✔ a 1.6.1 full-provenance sent view remains verifiable after introducing the compact view (1038.8736ms)
✔ actual HTTP requests keep four observed modules after a later large history read and preserve all source bytes (7793.5797ms)
ℹ tests 46
ℹ suites 0
ℹ pass 42
ℹ fail 4
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 85127.5674
✖ failing tests:
test at tests\reading-reliability.test.mjs:51:3
✖ chat-completions: complete actual Unicode/CRLF pages and small source survive the next bounded wire request (6433.4656ms)
AssertionError [ERR_ASSERTION]: Expected values to be strictly equal:
+ actual - expected
+ 'paused'
- 'completed'
at TestContext.<anonymous> (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/tests/reading-reliability.test.mjs:81:12)
at runNextTicks (node:internal/process/task_queues:65:5)
at process.processTimers (node:internal/timers:615:9)
at async Test.run (node:internal/test_runner/test:1409:7)
at async Test.processPendingSubtests (node:internal/test_runner/test:974:7) {
generatedMessage: true,
code: 'ERR_ASSERTION',
actual: 'paused',
expected: 'completed',
operator: 'strictEqual',
diff: 'simple'
}
test at tests\reading-reliability.test.mjs:51:3
✖ responses: complete actual Unicode/CRLF pages and small source survive the next bounded wire request (7626.023ms)
AssertionError [ERR_ASSERTION]: Expected values to be strictly equal:
+ actual - expected
+ 'paused'
- 'completed'
at TestContext.<anonymous> (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/tests/reading-reliability.test.mjs:81:12)
at runNextTicks (node:internal/process/task_queues:65:5)
at process.processTimers (node:internal/timers:615:9)
at async Test.run (node:internal/test_runner/test:1409:7)
at async Test.processPendingSubtests (node:internal/test_runner/test:974:7) {
generatedMessage: true,
code: 'ERR_ASSERTION',
actual: 'paused',
expected: 'completed',
operator: 'strictEqual',
diff: 'simple'
}
test at tests\reading-reliability.test.mjs:51:3
✖ anthropic-messages: complete actual Unicode/CRLF pages and small source survive the next bounded wire request (5372.4752ms)
AssertionError [ERR_ASSERTION]: Expected values to be strictly equal:
+ actual - expected
+ 'paused'
- 'completed'
at TestContext.<anonymous> (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/tests/reading-reliability.test.mjs:81:12)
at runNextTicks (node:internal/process/task_queues:65:5)
at process.processTimers (node:internal/timers:615:9)
at async Test.run (node:internal/test_runner/test:1409:7)
at async Test.processPendingSubtests (node:internal/test_runner/test:974:7) {
generatedMessage: true,
code: 'ERR_ASSERTION',
actual: 'paused',
expected: 'completed',
operator: 'strictEqual',
diff: 'simple'
}
test at tests\working-set.test.mjs:167:1
✖ actual HTTP file pagination advances over a long file with each returned page visible in the next request (8186.3182ms)
AssertionError [ERR_ASSERTION]: null
+ actual - expected
+ 'paused'
- 'completed'
at TestContext.<anonymous> (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/tests/working-set.test.mjs:185:10)
at async Test.run (node:internal/test_runner/test:1409:7)
at async Test.processPendingSubtests (node:internal/test_runner/test:974:7) {
generatedMessage: false,
code: 'ERR_ASSERTION',
actual: 'paused',
expected: 'completed',
operator: 'strictEqual',
diff: 'simple'
}
@@ -0,0 +1,107 @@
✔ Messages builds native system/tool history, bounds images, and seals a distinct mixed-pool variant (6.5484ms)
✔ Messages URL and native parameter contracts preserve other protocol paths and hard output budget (0.5985ms)
✔ native usage includes both cache buckets once, retains raw and rejects invalid counters (0.1786ms)
✔ real HTTP Messages JSON executes one tool, retains opaque history, actual URL and cache accounting (4313.6667ms)
✔ real HTTP Messages SSE executes one tool, retains opaque history, actual URL and cache accounting (3447.717ms)
✔ Messages api-key authentication is explicit; mismatched model remains blocked with original usage (6.1309ms)
✔ Messages bearer authentication is explicit; mismatched model remains blocked with original usage (4.5811ms)
✔ Messages redirect is not followed, either credential stays at the original server (3.6997ms)
✔ strict native SSE reads the nested start model and retains mismatch source without retries (3.0908ms)
✔ a closed native tool block cannot execute before message_stop, then executes exactly once (3213.8544ms)
✔ mixed Chat/Messages pool seals both variants and preserves one completed tool on failover (4246.1155ms)
✔ native JSON rejects incomplete stops, duplicate tools and hosted content (0.353ms)
▶ strict SSE validates all blocks, terminal, event names, UTF8 and tool JSON before dispatch
✔ reject native stream 0 (1.0888ms)
✔ reject native stream 1 (0.2255ms)
✔ reject native stream 2 (0.2871ms)
✔ reject native stream 3 (0.1879ms)
✔ reject native stream 4 (0.1582ms)
✔ reject native stream 5 (0.1619ms)
✔ reject native stream 6 (0.1227ms)
✔ reject native stream 7 (0.1105ms)
✔ reject native stream 8 (0.1843ms)
✔ strict SSE validates all blocks, terminal, event names, UTF8 and tool JSON before dispatch (3.1428ms)
✔ upstream SSE error retains original type/request ID and initial usage, never retries (4.7539ms)
✔ stop cancels the owned stream and retains partial consumption without running tool or replaying (61.3452ms)
✔ saved Messages import preserves auth choice, rejects endpoint duplication, and keeps source file untouched (958.0128ms)
✔ input that fits the engine window but exceeds remaining credit diagnoses the credit limit (0.5363ms)
✔ window overflow and historical pauses retain distinct and honest diagnoses (0.0809ms)
✔ R observed reservation shortage keeps the window-fit diagnosis separate from Run/Task credit (0.093ms)
✔ a budget pause persists concrete text/tool/history/image breakdown and never dispatches an oversized request (4056.5419ms)
✔ explicit frozen preview resumes the same Task in a new Run, preserves completed tools and does not change global config or selected model (4375.4502ms)
✔ insufficient budgets, silently clamped capacity and arbitrary config fields reject without a physical request or mutating task budget (2918.8302ms)
✔ changed session, changed registry, expiry and altered frozen preview all require a new explicit preview (3588.3646ms)
✔ preview preserves unknown invocation status and increasing Task allowance is committed only with confirmed new Run (5129.1533ms)
✔ saved engine budgets verify all Chat/Responses route variants and reject a mismatched output allowance (1075.63ms)
✔ owned CC Switch SQLite imports the selected TOML Responses wire_api and leaves unknown clients untouched (1455.6215ms)
✔ Responses executes one original tool through Runtime, carries encrypted Items, reports raw usage and seals exact wire bytes (3058.5134ms)
✔ mixed-protocol pool keeps both actual variants and changes route on 429 without re-executing completed tools (3965.986ms)
▶ Responses incomplete, failed, missing call identity and unsupported hosted outputs never dispatch a tool
✔ Reject {"body":{"id":"resp_owned","status":"incomplete","output":[{"type":"function_call","id":"fc_partial" (2033.5475ms)
✔ Reject {"body":{"id":"resp_owned","status":"failed","output":[{"type":"function_call","id":"fc_failed","cal (1919.3653ms)
✔ Reject {"body":{"id":"resp_owned","status":"completed","output":[{"type":"function_call","id":"fc_bad","cal (2027.8655ms)
✔ Reject {"body":{"id":"resp_owned","status":"completed","output":[{"type":"computer_call","id":"not-supporte (1631.6839ms)
✔ Responses incomplete, failed, missing call identity and unsupported hosted outputs never dispatch a tool (7613.4068ms)
▶ invalid JSON in a nonstream Chat or Responses tool group prevents every tool and pool replay
✔ chat-completions (1451.0538ms)
✔ responses (1388.3246ms)
✔ invalid JSON in a nonstream Chat or Responses tool group prevents every tool and pool replay (2839.7747ms)
✔ wire adapter preserves tool IDs but never passes Responses-private Items to Chat or another endpoint (0.2109ms)
✔ connection JSON import accepts explicit Responses protocol and continues to reject unknown protocols without reading client sessions (618.8068ms)
✔ reservation uses actual known usage from the same route, never lowers estimates or rewrites original receipts (1.5861ms)
✔ reservation verification binds original Task/Run consumption and rejects changed usage or foreign provenance (0.4653ms)
✔ known high input calibrates the next request and stops before dispatch while preserving the completed read (1951.0713ms)
✔ one reported overrun stops legacy requests and keeps actual use and excess (1509.2561ms)
✔ one reported overrun stops pool requests and keeps actual use and excess (2153.2021ms)
✔ one reported overrun stops connection requests and keeps actual use and excess (2059.8224ms)
✔ reported usage below the limit prevents an unaffordable legacy retry before another dispatch (1713.5456ms)
✔ reported usage below the limit prevents an unaffordable pool retry before another dispatch (2060.3406ms)
✔ native bridge seals calibrated refusal with original consumption and preserves versioned request verification (12821.7929ms)
✔ actual HTTP requests reuse only this project source-checked memory as redacted data and preserve sent evidence after archive/restart (1701.3743ms)
✔ actual HTTP request excludes archived, timed-out and foreign-project memories (2694.8128ms)
✔ actual next HTTP request removes expired memory after a tool observes a source change without replaying that tool (2237.4216ms)
✔ incomplete inventory and excluded-file SHA changes cannot enter actual HTTP memory context or inspect_session snapshots (2585.9887ms)
✔ corrupt original artifact expires its memory and cannot reach an HTTP model request (1997.7985ms)
✔ memory JSON is bounded in escaped UTF-8 bytes and entry count, skips oversized entries and retains complete multilingual values (4106.7101ms)
✔ optional memory yields to the complete goal and original code evidence under the whole-request budget (1376.0617ms)
✔ changed sealed memory artifacts or binding fields fail verification while the original request remains immutable (2487.6342ms)
✔ a malformed project declaration omits optional memory while the repair Task still reaches the owned HTTP service (2576.1328ms)
✔ an expiration persistence failure prevents dispatch and is not hidden as an optional scan failure (1944.4406ms)
✔ continuation reads advance through historical pages without a false pause while repeated current-Run pages still pause (1.8851ms)
✔ a new tool page replaces old latest-feedback space and still obeys its escaped protocol budget (0.4571ms)
✔ budgeted artifact pages count escaped protocol bytes and reconstruct original multilingual JSON without losing characters (9.0225ms)
✔ known artifact coverage detects semantic repeats despite changing limits, and advancing pages or direct evidence restore progress (0.5303ms)
✔ actual large artifact reads are present in the next bounded HTTP request and next_offset progresses without redoing tools (4568.8846ms)
✔ a real repeated artifact loop pauses the same Task after recorded results, survives reload and only resumes explicitly (9154.0295ms)
✔ actual HTTP continuation after reload sends distinct historically covered pages without false no_progress or replay (9290.3679ms)
✔ continuations distinguish unread source, later original line ranges and already returned result wrappers (1391.3603ms)
✔ pure-file batch returns unused small-page space while mixed and historical allocation stays conservative (0.3365ms)
✔ observed upstream high input pauses before a new request, preserves reported usage and distinguishes the limiting ledger (1119.5178ms)
✔ chat-completions: complete actual Unicode/CRLF pages and small source survive the next bounded wire request (4489.4919ms)
✔ responses: complete actual Unicode/CRLF pages and small source survive the next bounded wire request (5136.0782ms)
✔ anthropic-messages: complete actual Unicode/CRLF pages and small source survive the next bounded wire request (4040.5572ms)
✔ a real two-reply failure replay reaches request three at 16K with complete goal, original sources and sealed tool evidence (10517.5929ms)
✔ request deduplication retains long original intent, extra requirements, constraints and acceptance without changing the stored contract (2958.3216ms)
✔ only the routine context avoids plan echoes; diagnostic API and requests without an included plan keep original observations (826.171ms)
✔ small code and actual failed command stay visible when their original multi-tool group cannot fit (3228.0359ms)
✔ mixed group retains six distinct small file observations with exact multilingual text and bounded request bytes (3286.0909ms)
✔ working observations are bound to the current Task and stop treating a pre-application file revision as current evidence (2421.2836ms)
✔ a sealed observation rejects changed or missing original result artifacts instead of recreating evidence (1558.8453ms)
✔ a real passed-check observation is excluded after a source edit invalidates the check, while its historical sent evidence remains valid (5042.6206ms)
✔ budgeted file pages reconstruct UTF-8 text including long single lines and surrogate boundaries (4.7148ms)
✔ actual HTTP multi-tool reads survive an oversized runtime snapshot and the sealed evidence survives reopening (11973.8585ms)
✔ actual HTTP file pagination advances over a long file with each returned page visible in the next request (6994.9477ms)
✔ later mixed history pages cannot displace the actual module sources and complete requirements (6745.4624ms)
✔ a later narrow page of the same file revision cannot evict the wider actual observation (3898.4259ms)
✔ six small source files stay complete across metadata turns even when mandatory evidence exceeds the soft target (3601.4413ms)
✔ historical read-artifact navigation points to the original artifact and next page, never its result wrapper (966.1327ms)
✔ a 1.6.1 full-provenance sent view remains verifiable after introducing the compact view (1104.2318ms)
✔ actual HTTP requests keep four observed modules after a later large history read and preserve all source bytes (9402.6698ms)
ℹ tests 96
ℹ suites 0
ℹ pass 96
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 137955.7445
@@ -0,0 +1,111 @@
✔ input that fits the engine window but exceeds remaining credit diagnoses the credit limit (0.5051ms)
✔ window overflow and historical pauses retain distinct and honest diagnoses (0.0846ms)
✔ continuation reads advance through historical pages without a false pause while repeated current-Run pages still pause (1.8791ms)
✔ a new tool page replaces old latest-feedback space and still obeys its escaped protocol budget (0.4529ms)
✔ budgeted artifact pages count escaped protocol bytes and reconstruct original multilingual JSON without losing characters (9.1764ms)
✔ known artifact coverage detects semantic repeats despite changing limits, and advancing pages or direct evidence restore progress (0.347ms)
✖ actual large artifact reads are present in the next bounded HTTP request and next_offset progresses without redoing tools (5294.1931ms)
✔ a real repeated artifact loop pauses the same Task after recorded results, survives reload and only resumes explicitly (5228.3237ms)
✔ actual HTTP continuation after reload sends distinct historically covered pages without false no_progress or replay (4499.8861ms)
✖ a real two-reply failure replay reaches request three at 16K with complete goal, original sources and sealed tool evidence (4956.0812ms)
✔ request deduplication retains long original intent, extra requirements, constraints and acceptance without changing the stored contract (1487.7346ms)
✔ only the routine context avoids plan echoes; diagnostic API and requests without an included plan keep original observations (489.9709ms)
✔ small code and actual failed command stay visible when their original multi-tool group cannot fit (1588.7915ms)
✔ mixed group retains six distinct small file observations with exact multilingual text and bounded request bytes (1926.1681ms)
✔ working observations are bound to the current Task and stop treating a pre-application file revision as current evidence (1750.2454ms)
✔ a sealed observation rejects changed or missing original result artifacts instead of recreating evidence (1232.0396ms)
✔ a real passed-check observation is excluded after a source edit invalidates the check, while its historical sent evidence remains valid (4412.558ms)
✔ budgeted file pages reconstruct UTF-8 text including long single lines and surrogate boundaries (4.2331ms)
✔ actual HTTP multi-tool reads survive an oversized runtime snapshot and the sealed evidence survives reopening (6777.866ms)
✔ actual HTTP file pagination advances over a long file with each returned page visible in the next request (2852.6515ms)
✔ later mixed history pages cannot displace the actual module sources and complete requirements (4169.3804ms)
✔ a later narrow page of the same file revision cannot evict the wider actual observation (2328.55ms)
✔ six small source files stay complete across metadata turns even when mandatory evidence exceeds the soft target (2436.9966ms)
✔ historical read-artifact navigation points to the original artifact and next page, never its result wrapper (1030.0164ms)
✖ a 1.6.1 full-provenance sent view remains verifiable after introducing the compact view (927.5454ms)
✖ actual HTTP requests keep four observed modules after a later large history read and preserve all source bytes (9519.075ms)
ℹ tests 26
ℹ suites 0
ℹ pass 22
ℹ fail 4
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 48315.6474
✖ failing tests:
test at tests\read-budget.test.mjs:73:1
✖ actual large artifact reads are present in the next bounded HTTP request and next_offset progresses without redoing tools (5294.1931ms)
AssertionError [ERR_ASSERTION]: {"failure":{"id":"9e6d310c-37ec-4f38-95b5-49a6d7acea32","source":"model_connection","origin":"upstream","stage":"model_request","kind":"service","httpStatus":500,"endpoint":"http://127.0.0.1:52362/v1","model":"page-budget-fixture","route":"/chat/completions","attempts":3,"occurredAt":"2026-11-03T16:31:32.819Z","retryable":true,"title":"模型服务暂时无法处理请求","upstream":{"schema":1,"scope":"http-body","httpStatus":500,"code":null,"type":null,"message":"Newest actual result must be sent, not archived before the model sees it","param":null,"reason":null,"requestId":null,"responseId":null,"eventType":null,"headers":{"content-type":"application/json","date":"Tue, 03 Nov 2026 16:31:32 GMT"},"retryAfter":null,"retryAfterMs":null,"rawPreview":"{\"error\":{\"message\":\"Newest actual result must be sent, not archived before the model sees it\"}}","previewTruncated":false,"capture":{"scope":"http-body","receivedBytes":96,"capturedBytes":96,"storedBytes":96,"retention":"artifact","complete":true,"truncated":false,"interrupted":false,"capturedRange":[0,96],"redacted":false,"sha256":"bc1dd20cfb838626ba925ecf33e0b0ee15648ee4b11c588226aac3517a6f7a2d"},"artifactId":"7eb611e5-cd1f-42b5-928f-ae75b02cfb84","artifactSha256":"bc1dd20cfb838626ba925ecf33e0b0ee15648ee4b11c588226aac3517a6f7a2d","artifactScope":"http-body","sessionId":"8cdd5f07-5a9d-4b3a-abb4-8b635fcee700"},"upstreamRequestId":null,"code":null,"type":null,"limitReason":null,"wireAttemptId":"b4d55d94-626f-4c82-a3e2-4acfcfc2b46c","message":"模型服务暂时无法处理请求(HTTP 500)。 Newest actual result must be sent, not archived before the model sees it","detail":"Newest actual result must be sent, not archived before the model sees it","suggestion":"稍后重试,或检查模型服务状态。","requestUrl":"http://127.0.0.1:52362/v1/chat/completions","responseUrl":"http://127.0.0.1:52362/v1/chat/completions","redirected":false,"redirectMode":"manual","retryDecision":"attempt_limit","remainingWaitBudgetMs":297260,"runId":"d67c03a1-2f1e-4f09-b896-4861a413ce05","messageId":"e5569f34-4897-48f3-bc72-dd6c0ec6d328"},"requests":4,"pages":[]}
+ actual - expected
+ 'failed'
- 'completed'
at TestContext.<anonymous> (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/tests/read-budget.test.mjs:94:10)
at runNextTicks (node:internal/process/task_queues:65:5)
at process.processTimers (node:internal/timers:615:9)
at async Test.run (node:internal/test_runner/test:1409:7)
at async Test.processPendingSubtests (node:internal/test_runner/test:974:7) {
generatedMessage: false,
code: 'ERR_ASSERTION',
actual: 'failed',
expected: 'completed',
operator: 'strictEqual',
diff: 'simple'
}
test at tests\task-context.test.mjs:19:1
✖ a real two-reply failure replay reaches request three at 16K with complete goal, original sources and sealed tool evidence (4956.0812ms)
AssertionError [ERR_ASSERTION]: {"schema":1,"reason":"context_limit","createdAt":"2026-11-03T16:31:32.650Z","taskId":"28beecb1-6857-4bc1-b343-d99cf8ba14f1","runId":"7b04bf0f-3984-4957-a312-93a17e480928","contextLimit":16000,"outputReservation":2048,"safetyTokens":1600,"inputBudget":12352,"targetInputTokens":12352,"runRemainingTokens":89664,"taskRemainingTokens":89664,"limitingReason":"context_limit","requiredInputTokens":13343,"imageReservationTokens":0,"imageEstimate":null,"breakdown":{"transportAndImageFramingTokens":36,"goalAndContextTextTokens":5053,"toolDefinitionTokens":4622,"historyAndEvidenceTokens":3632,"imageReservationTokens":0},"runUsage":{"requests":2,"inputTokens":246,"outputTokens":90,"totalTokens":336,"unknownRequests":0,"estimatedUnknownTokens":0},"taskUsage":{"requests":2,"inputTokens":246,"outputTokens":90,"totalTokens":336,"unknownRequests":0,"estimatedUnknownTokens":0},"tokenAccounting":{"schemaVersion":1,"singleInFlightResponseMayExceed":true,"run":{"maximumTokens":90000,"reportedTokens":336,"unknownRequests":0,"estimatedUnknownTokens":0,"accountedTokens":336,"remainingTokens":89664,"exceededTokens":0},"task":{"maximumTokens":90000,"reportedTokens":336,"unknownRequests":0,"estimatedUnknownTokens":0,"accountedTokens":336,"remainingTokens":89664,"exceededTokens":0}},"runBudget":{"maxSteps":8,"maxDurationMs":600000,"maxTokens":90000,"maxTaskTokens":90000},"taskBudget":{"maxTokens":90000},"note":"本地预留与上游报告用量分开记录。单个在途响应可能超过预估;超额后停止后续请求,未知用量保持未知。"}
2 !== 3
at TestContext.<anonymous> (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/tests/task-context.test.mjs:44:10)
at async Test.run (node:internal/test_runner/test:1409:7)
at async startSubtestAfterBootstrap (node:internal/test_runner/harness:387:3) {
generatedMessage: false,
code: 'ERR_ASSERTION',
actual: 2,
expected: 3,
operator: 'strictEqual',
diff: 'simple'
}
test at tests\working-set.test.mjs:275:1
✖ a 1.6.1 full-provenance sent view remains verifiable after introducing the compact view (927.5454ms)
AssertionError [ERR_ASSERTION]: Missing expected exception.
at TestContext.<anonymous> (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/tests/working-set.test.mjs:287:10)
at Test.runInAsyncScope (node:async_hooks:227:14)
at Test.run (node:internal/test_runner/test:1402:25)
at Test.processPendingSubtests (node:internal/test_runner/test:974:18)
at Test.postRun (node:internal/test_runner/test:1542:19)
at Test.run (node:internal/test_runner/test:1467:12)
at async Test.processPendingSubtests (node:internal/test_runner/test:974:7) {
generatedMessage: false,
code: 'ERR_ASSERTION',
actual: undefined,
expected: /模型预算预留与原估算\/用量\/实际线路不一致/,
operator: 'throws',
diff: 'simple'
}
test at tests\working-set.test.mjs:298:1
✖ actual HTTP requests keep four observed modules after a later large history read and preserve all source bytes (9519.075ms)
AssertionError [ERR_ASSERTION]: {"id":"60fc5104-6308-4fdb-b8e0-83d60d028154","source":"model_connection","origin":"upstream","stage":"model_request","kind":"service","httpStatus":500,"endpoint":"http://127.0.0.1:62976/v1","model":"working-retention-http-fixture","route":"/chat/completions","attempts":3,"occurredAt":"2026-11-03T16:32:15.746Z","retryable":true,"title":"模型服务暂时无法处理请求","upstream":{"schema":1,"scope":"http-body","httpStatus":500,"code":null,"type":null,"message":"Lost actual module after historical read: src/normalize.mjs\n+ actual - expected\n\n+ undefined\n- 'export const normalize = value => value; // observed 雪🙂\\n'\n","param":null,"reason":null,"requestId":null,"responseId":null,"eventType":null,"headers":{"content-type":"application/json","date":"Tue, 03 Nov 2026 16:32:15 GMT"},"retryAfter":null,"retryAfterMs":null,"rawPreview":"{\"error\":{\"message\":\"Lost actual module after historical read: src/normalize.mjs\\n+ actual - expected\\n\\n+ undefined\\n- 'export const normalize = value => value; // observed 雪🙂\\\\n'\\n\"}}","previewTruncated":false,"capture":{"scope":"http-body","receivedBytes":190,"capturedBytes":190,"storedBytes":190,"retention":"artifact","complete":true,"truncated":false,"interrupted":false,"capturedRange":[0,190],"redacted":false,"sha256":"1776104235024728aa9e050e80d928223e48a4278ec6ed08abab008cc35fa694"},"artifactId":"4d2980a3-cb5b-4d1b-9123-092b05bebd38","artifactSha256":"1776104235024728aa9e050e80d928223e48a4278ec6ed08abab008cc35fa694","artifactScope":"http-body","sessionId":"11c46ef1-f114-4926-9629-6cbf2abf82ec"},"upstreamRequestId":null,"code":null,"type":null,"limitReason":null,"wireAttemptId":"0716b9a5-00a6-48bb-8152-96de1e606ba9","message":"模型服务暂时无法处理请求(HTTP 500)。 Lost actual module after historical read: src/normalize.mjs\n+ actual - expected\n\n+ undefined\n- 'export const normalize = value => value; // observed 雪🙂\\n'\n","detail":"Lost actual module after historical read: src/normalize.mjs\n+ actual - expected\n\n+ undefined\n- 'export const normalize = value => value; // observed 雪🙂\\n'\n","suggestion":"稍后重试,或检查模型服务状态。","requestUrl":"http://127.0.0.1:62976/v1/chat/completions","responseUrl":"http://127.0.0.1:62976/v1/chat/completions","redirected":false,"redirectMode":"manual","retryDecision":"attempt_limit","remainingWaitBudgetMs":297427,"runId":"34db6ae3-fca1-4625-ad2b-703dc77a2bd0","messageId":"cd1efb50-ccae-4a58-bdd7-c097c924d172"}
+ actual - expected
+ 'failed'
- 'completed'
at TestContext.<anonymous> (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/tests/working-set.test.mjs:335:10)
at async Test.run (node:internal/test_runner/test:1409:7)
at async Test.processPendingSubtests (node:internal/test_runner/test:974:7) {
generatedMessage: false,
code: 'ERR_ASSERTION',
actual: 'failed',
expected: 'completed',
operator: 'strictEqual',
diff: 'simple'
}
@@ -0,0 +1,49 @@
✔ continuations distinguish unread source, later original line ranges and already returned result wrappers (1045.7177ms)
✔ pure-file batch returns unused small-page space while mixed and historical allocation stays conservative (0.2933ms)
✔ observed upstream high input pauses before a new request, preserves reported usage and distinguishes the limiting ledger (953.6998ms)
✔ chat-completions: complete actual Unicode/CRLF pages and small source survive the next bounded wire request (6089.461ms)
✔ responses: complete actual Unicode/CRLF pages and small source survive the next bounded wire request (6364.6585ms)
✔ anthropic-messages: complete actual Unicode/CRLF pages and small source survive the next bounded wire request (7978.6616ms)
✔ small code and actual failed command stay visible when their original multi-tool group cannot fit (2508.6503ms)
✔ mixed group retains six distinct small file observations with exact multilingual text and bounded request bytes (1947.5017ms)
✔ working observations are bound to the current Task and stop treating a pre-application file revision as current evidence (1788.2467ms)
✔ a sealed observation rejects changed or missing original result artifacts instead of recreating evidence (1220.9115ms)
✔ a real passed-check observation is excluded after a source edit invalidates the check, while its historical sent evidence remains valid (5515.1539ms)
✔ budgeted file pages reconstruct UTF-8 text including long single lines and surrogate boundaries (3.9342ms)
✔ actual HTTP multi-tool reads survive an oversized runtime snapshot and the sealed evidence survives reopening (11012.9135ms)
✖ actual HTTP file pagination advances over a long file with each returned page visible in the next request (8045.7707ms)
✔ later mixed history pages cannot displace the actual module sources and complete requirements (4533.8981ms)
✔ a later narrow page of the same file revision cannot evict the wider actual observation (2278.9637ms)
✔ six small source files stay complete across metadata turns even when mandatory evidence exceeds the soft target (2428.4988ms)
✔ historical read-artifact navigation points to the original artifact and next page, never its result wrapper (949.7264ms)
✔ a 1.6.1 full-provenance sent view remains verifiable after introducing the compact view (1135.4206ms)
✔ actual HTTP requests keep four observed modules after a later large history read and preserve all source bytes (8558.3936ms)
ℹ tests 20
ℹ suites 0
ℹ pass 19
ℹ fail 1
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 52251.0733
✖ failing tests:
test at tests\working-set.test.mjs:167:1
✖ actual HTTP file pagination advances over a long file with each returned page visible in the next request (8045.7707ms)
AssertionError [ERR_ASSERTION]: null
+ actual - expected
+ 'paused'
- 'completed'
at TestContext.<anonymous> (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/tests/working-set.test.mjs:185:10)
at async Test.run (node:internal/test_runner/test:1409:7)
at async Test.processPendingSubtests (node:internal/test_runner/test:974:7) {
generatedMessage: false,
code: 'ERR_ASSERTION',
actual: 'paused',
expected: 'completed',
operator: 'strictEqual',
diff: 'simple'
}
@@ -0,0 +1,25 @@
{
"success": true,
"version": "1.24.6",
"pid": 36552,
"projects": 63,
"sessions": 79,
"tables": {
"events": 13279,
"metadata": 7,
"outbox": 13279,
"projects": 63,
"records": 6262,
"sessions": 79,
"sqlite_sequence": 1
},
"runtimeRoot": "C:\\Users\\wangy\\AppData\\Local\\Programs\\AgentWorkbench\\versions\\Agent-Workbench-1.24.6-win-x64-20261103151145294",
"sourceCommit": "5dcd6d7c3f4f465071ae1156bb1e0bf5a28f48e2",
"hasKey": true,
"active": [],
"mode": "before",
"startedAt": "2026-11-03T17:11:28.612Z",
"finishedAt": "2026-11-03T17:11:36.551Z",
"realModelRequests": 0,
"scope": "Public API/full original row/export/artifact hashes; installer retains original running host and Provider. No credential extraction or task replay."
}
File diff suppressed because one or more lines are too long
@@ -0,0 +1,401 @@
{
"success": true,
"version": "1.24.7",
"pid": 36552,
"runtimeRoot": "C:\\Users\\wangy\\AppData\\Local\\Programs\\AgentWorkbench\\versions\\Agent-Workbench-1.24.7-win-x64-20261103164618190",
"projects": 63,
"sessions": 79,
"hasKey": true,
"retention": {
"originalRows": "exact",
"originalExports": "exact",
"originalArtifactBytes": "exact",
"originalModelCenter": "exact"
},
"pointer": {
"schema": 1,
"app": "agent-workbench-installation",
"installId": "74c11bdd-4895-459b-8403-097f79980b41",
"revision": 23,
"dataRoot": "C:\\Users\\wangy\\AppData\\Local\\Packages\\OpenAI.Codex_2p2nqsd0c76g0\\LocalCache\\Local\\AgentWorkbench",
"port": 4318,
"active": {
"directory": "Agent-Workbench-1.24.7-win-x64-20261103164618190",
"version": "1.24.7",
"manifestSha256": "03683cabd6b88cabdb3eb4cd159ffbabdf1b37e1576354ec3d6433d4ca773be9",
"runtimeIdentity": "212aba55ceb5eb08bb7985dd346bfb3967f5c7a36b2bc848b7e149adb50edf85"
},
"previous": {
"directory": "Agent-Workbench-1.24.6-win-x64-20261103151145294",
"version": "1.24.6",
"manifestSha256": "64aa8b2ae90df30eb33b9d2f60b458c3fcf61de7cbc778ac2d08fd0f7437127d",
"runtimeIdentity": "2c662dee28c0c61b57e2f77dbb546667af7b6938aaf848e5a6facde973cad2e1"
},
"activatedAt": "2026-11-03T17:26:38.020Z"
},
"operation": {
"schema": 1,
"id": "f629056a0fd895cf154e56aab51b2c042680d455eddb9a70824cbe6bfbde6c83",
"installRoot": "C:\\Users\\wangy\\AppData\\Local\\Programs\\AgentWorkbench",
"dataRoot": "C:\\Users\\wangy\\AppData\\Local\\Packages\\OpenAI.Codex_2p2nqsd0c76g0\\LocalCache\\Local\\AgentWorkbench",
"command": "activate",
"expectedManifestSha256": "03683cabd6b88cabdb3eb4cd159ffbabdf1b37e1576354ec3d6433d4ca773be9",
"startedAt": "2026-11-03T17:12:08.648Z",
"sequence": 1579,
"history": [
{
"stage": "preparing",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:12:09.216Z",
"finishedAt": "2026-11-03T17:12:09.221Z",
"elapsedMs": 5
},
{
"stage": "verify-current",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:12:09.221Z",
"finishedAt": "2026-11-03T17:12:09.249Z",
"elapsedMs": 28
},
{
"stage": "verify-current",
"scope": "full-manifest",
"counts": {
"completedFiles": 10944,
"totalFiles": 10944,
"completedBytes": 657071308,
"totalBytes": 657071308,
"currentPath": "scripts/upgrade-running.mjs"
},
"startedAt": "2026-11-03T17:12:09.249Z",
"finishedAt": "2026-11-03T17:13:14.482Z",
"elapsedMs": 65233
},
{
"stage": "verify-current",
"scope": "directory-inventory",
"counts": {
"completedFiles": 10945,
"totalFiles": 10945,
"completedBytes": null,
"totalBytes": null,
"currentPath": "THIRD_PARTY_NOTICES.txt"
},
"startedAt": "2026-11-03T17:13:14.482Z",
"finishedAt": "2026-11-03T17:13:56.133Z",
"elapsedMs": 41651
},
{
"stage": "verify-current",
"scope": "loaded-inputs",
"counts": {
"completedFiles": 22,
"totalFiles": 22,
"completedBytes": 9306746,
"totalBytes": 9306746,
"currentPath": "dist-server/server.mjs"
},
"startedAt": "2026-11-03T17:13:56.133Z",
"finishedAt": "2026-11-03T17:13:56.266Z",
"elapsedMs": 134
},
{
"stage": "inspect-root",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:13:56.267Z",
"finishedAt": "2026-11-03T17:13:56.297Z",
"elapsedMs": 31
},
{
"stage": "inspect-host",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:13:56.297Z",
"finishedAt": "2026-11-03T17:13:59.827Z",
"elapsedMs": 3530
},
{
"stage": "preparing",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:14:35.216Z",
"finishedAt": "2026-11-03T17:14:35.222Z",
"elapsedMs": 6
},
{
"stage": "verify-candidate",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:14:35.222Z",
"finishedAt": "2026-11-03T17:14:35.252Z",
"elapsedMs": 30
},
{
"stage": "verify-candidate",
"scope": "full-manifest",
"counts": {
"completedFiles": 10949,
"totalFiles": 10949,
"completedBytes": 657252428,
"totalBytes": 657252428,
"currentPath": "scripts/upgrade-running.mjs"
},
"startedAt": "2026-11-03T17:14:35.252Z",
"finishedAt": "2026-11-03T17:17:10.939Z",
"elapsedMs": 155688
},
{
"stage": "verify-candidate",
"scope": "directory-inventory",
"counts": {
"completedFiles": 10950,
"totalFiles": 10950,
"completedBytes": null,
"totalBytes": null,
"currentPath": "THIRD_PARTY_NOTICES.txt"
},
"startedAt": "2026-11-03T17:17:10.939Z",
"finishedAt": "2026-11-03T17:17:51.843Z",
"elapsedMs": 40904
},
{
"stage": "verify-candidate",
"scope": "loaded-inputs",
"counts": {
"completedFiles": 22,
"totalFiles": 22,
"completedBytes": 9321617,
"totalBytes": 9321617,
"currentPath": "dist-server/server.mjs"
},
"startedAt": "2026-11-03T17:17:51.843Z",
"finishedAt": "2026-11-03T17:17:53.084Z",
"elapsedMs": 1241
},
{
"stage": "verify-current",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:17:53.084Z",
"finishedAt": "2026-11-03T17:17:53.121Z",
"elapsedMs": 37
},
{
"stage": "verify-current",
"scope": "full-manifest",
"counts": {
"completedFiles": 10944,
"totalFiles": 10944,
"completedBytes": 657071308,
"totalBytes": 657071308,
"currentPath": "scripts/upgrade-running.mjs"
},
"startedAt": "2026-11-03T17:17:53.121Z",
"finishedAt": "2026-11-03T17:20:29.767Z",
"elapsedMs": 156647
},
{
"stage": "verify-current",
"scope": "directory-inventory",
"counts": {
"completedFiles": 10945,
"totalFiles": 10945,
"completedBytes": null,
"totalBytes": null,
"currentPath": "THIRD_PARTY_NOTICES.txt"
},
"startedAt": "2026-11-03T17:20:29.767Z",
"finishedAt": "2026-11-03T17:21:10.694Z",
"elapsedMs": 40927
},
{
"stage": "verify-current",
"scope": "loaded-inputs",
"counts": {
"completedFiles": 22,
"totalFiles": 22,
"completedBytes": 9306746,
"totalBytes": 9306746,
"currentPath": "dist-server/server.mjs"
},
"startedAt": "2026-11-03T17:21:10.694Z",
"finishedAt": "2026-11-03T17:21:14.556Z",
"elapsedMs": 3862
},
{
"stage": "snapshot-data",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:21:14.556Z",
"finishedAt": "2026-11-03T17:21:37.148Z",
"elapsedMs": 22592
},
{
"stage": "retain-host",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:21:37.148Z",
"finishedAt": "2026-11-03T17:21:41.330Z",
"elapsedMs": 4182
},
{
"stage": "verify-data",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:21:41.330Z",
"finishedAt": "2026-11-03T17:21:48.048Z",
"elapsedMs": 6717
},
{
"stage": "verify-current",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:21:48.048Z",
"finishedAt": "2026-11-03T17:21:48.075Z",
"elapsedMs": 27
},
{
"stage": "verify-current",
"scope": "full-manifest",
"counts": {
"completedFiles": 10944,
"totalFiles": 10944,
"completedBytes": 657071308,
"totalBytes": 657071308,
"currentPath": "scripts/upgrade-running.mjs"
},
"startedAt": "2026-11-03T17:21:48.075Z",
"finishedAt": "2026-11-03T17:22:49.894Z",
"elapsedMs": 61818
},
{
"stage": "verify-current",
"scope": "directory-inventory",
"counts": {
"completedFiles": 10945,
"totalFiles": 10945,
"completedBytes": null,
"totalBytes": null,
"currentPath": "THIRD_PARTY_NOTICES.txt"
},
"startedAt": "2026-11-03T17:22:49.894Z",
"finishedAt": "2026-11-03T17:23:30.779Z",
"elapsedMs": 40885
},
{
"stage": "verify-current",
"scope": "loaded-inputs",
"counts": {
"completedFiles": 22,
"totalFiles": 22,
"completedBytes": 9306746,
"totalBytes": 9306746,
"currentPath": "dist-server/server.mjs"
},
"startedAt": "2026-11-03T17:23:30.779Z",
"finishedAt": "2026-11-03T17:23:30.909Z",
"elapsedMs": 130
},
{
"stage": "verify-candidate",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:23:30.909Z",
"finishedAt": "2026-11-03T17:23:30.932Z",
"elapsedMs": 23
},
{
"stage": "verify-candidate",
"scope": "full-manifest",
"counts": {
"completedFiles": 10949,
"totalFiles": 10949,
"completedBytes": 657252428,
"totalBytes": 657252428,
"currentPath": "scripts/upgrade-running.mjs"
},
"startedAt": "2026-11-03T17:23:30.932Z",
"finishedAt": "2026-11-03T17:25:56.919Z",
"elapsedMs": 145988
},
{
"stage": "verify-candidate",
"scope": "directory-inventory",
"counts": {
"completedFiles": 10950,
"totalFiles": 10950,
"completedBytes": null,
"totalBytes": null,
"currentPath": "THIRD_PARTY_NOTICES.txt"
},
"startedAt": "2026-11-03T17:25:56.919Z",
"finishedAt": "2026-11-03T17:26:37.873Z",
"elapsedMs": 40954
},
{
"stage": "verify-candidate",
"scope": "loaded-inputs",
"counts": {
"completedFiles": 22,
"totalFiles": 22,
"completedBytes": 9321617,
"totalBytes": 9321617,
"currentPath": "dist-server/server.mjs"
},
"startedAt": "2026-11-03T17:26:37.873Z",
"finishedAt": "2026-11-03T17:26:38.012Z",
"elapsedMs": 139
},
{
"stage": "publish-pointer",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:26:38.012Z",
"finishedAt": "2026-11-03T17:26:38.025Z",
"elapsedMs": 14
},
{
"stage": "verify-publication",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:26:38.025Z",
"finishedAt": "2026-11-03T17:26:38.073Z",
"elapsedMs": 48
}
],
"process": {
"pid": 15272,
"createdAt": "2026-11-03T17:14:33.493Z"
},
"candidate": {
"root": "C:\\Users\\wangy\\AppData\\Local\\Programs\\AgentWorkbench\\versions\\Agent-Workbench-1.24.7-win-x64-20261103164618190",
"version": "1.24.7",
"manifestSha256": "03683cabd6b88cabdb3eb4cd159ffbabdf1b37e1576354ec3d6433d4ca773be9",
"runtimeIdentity": "212aba55ceb5eb08bb7985dd346bfb3967f5c7a36b2bc848b7e149adb50edf85"
},
"state": "completed",
"stage": "completed",
"scope": null,
"counts": null,
"interruptible": false,
"label": "版本激活完成",
"detail": {
"activeVersion": "1.24.7",
"receiptFile": "C:\\Users\\wangy\\AppData\\Local\\Programs\\AgentWorkbench\\.installation\\operation-a9edbc42-e65b-44f3-bfed-a93e7a8324c4.json"
},
"stageStartedAt": "2026-11-03T17:26:38.073Z",
"updatedAt": "2026-11-03T17:26:38.073Z",
"elapsedMs": 869425
},
"desktopWindows": 1,
"inspectorClosed": true,
"sourceCommit": "5dcd6d7c3f4f465071ae1156bb1e0bf5a28f48e2",
"mode": "final",
"startedAt": "2026-11-03T17:30:47.536Z",
"finishedAt": "2026-11-03T17:31:04.150Z",
"realModelRequests": 0,
"scope": "Public API/full original row/export/artifact hashes; installer retains original running host and Provider. No credential extraction or task replay."
}
File diff suppressed because one or more lines are too long
@@ -0,0 +1,401 @@
{
"success": true,
"version": "1.24.7",
"pid": 36552,
"runtimeRoot": "C:\\Users\\wangy\\AppData\\Local\\Programs\\AgentWorkbench\\versions\\Agent-Workbench-1.24.7-win-x64-20261103164618190",
"projects": 63,
"sessions": 79,
"hasKey": true,
"retention": {
"originalRows": "exact",
"originalExports": "exact",
"originalArtifactBytes": "exact",
"originalModelCenter": "exact"
},
"pointer": {
"schema": 1,
"app": "agent-workbench-installation",
"installId": "74c11bdd-4895-459b-8403-097f79980b41",
"revision": 23,
"dataRoot": "C:\\Users\\wangy\\AppData\\Local\\Packages\\OpenAI.Codex_2p2nqsd0c76g0\\LocalCache\\Local\\AgentWorkbench",
"port": 4318,
"active": {
"directory": "Agent-Workbench-1.24.7-win-x64-20261103164618190",
"version": "1.24.7",
"manifestSha256": "03683cabd6b88cabdb3eb4cd159ffbabdf1b37e1576354ec3d6433d4ca773be9",
"runtimeIdentity": "212aba55ceb5eb08bb7985dd346bfb3967f5c7a36b2bc848b7e149adb50edf85"
},
"previous": {
"directory": "Agent-Workbench-1.24.6-win-x64-20261103151145294",
"version": "1.24.6",
"manifestSha256": "64aa8b2ae90df30eb33b9d2f60b458c3fcf61de7cbc778ac2d08fd0f7437127d",
"runtimeIdentity": "2c662dee28c0c61b57e2f77dbb546667af7b6938aaf848e5a6facde973cad2e1"
},
"activatedAt": "2026-11-03T17:26:38.020Z"
},
"operation": {
"schema": 1,
"id": "f629056a0fd895cf154e56aab51b2c042680d455eddb9a70824cbe6bfbde6c83",
"installRoot": "C:\\Users\\wangy\\AppData\\Local\\Programs\\AgentWorkbench",
"dataRoot": "C:\\Users\\wangy\\AppData\\Local\\Packages\\OpenAI.Codex_2p2nqsd0c76g0\\LocalCache\\Local\\AgentWorkbench",
"command": "activate",
"expectedManifestSha256": "03683cabd6b88cabdb3eb4cd159ffbabdf1b37e1576354ec3d6433d4ca773be9",
"startedAt": "2026-11-03T17:12:08.648Z",
"sequence": 1579,
"history": [
{
"stage": "preparing",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:12:09.216Z",
"finishedAt": "2026-11-03T17:12:09.221Z",
"elapsedMs": 5
},
{
"stage": "verify-current",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:12:09.221Z",
"finishedAt": "2026-11-03T17:12:09.249Z",
"elapsedMs": 28
},
{
"stage": "verify-current",
"scope": "full-manifest",
"counts": {
"completedFiles": 10944,
"totalFiles": 10944,
"completedBytes": 657071308,
"totalBytes": 657071308,
"currentPath": "scripts/upgrade-running.mjs"
},
"startedAt": "2026-11-03T17:12:09.249Z",
"finishedAt": "2026-11-03T17:13:14.482Z",
"elapsedMs": 65233
},
{
"stage": "verify-current",
"scope": "directory-inventory",
"counts": {
"completedFiles": 10945,
"totalFiles": 10945,
"completedBytes": null,
"totalBytes": null,
"currentPath": "THIRD_PARTY_NOTICES.txt"
},
"startedAt": "2026-11-03T17:13:14.482Z",
"finishedAt": "2026-11-03T17:13:56.133Z",
"elapsedMs": 41651
},
{
"stage": "verify-current",
"scope": "loaded-inputs",
"counts": {
"completedFiles": 22,
"totalFiles": 22,
"completedBytes": 9306746,
"totalBytes": 9306746,
"currentPath": "dist-server/server.mjs"
},
"startedAt": "2026-11-03T17:13:56.133Z",
"finishedAt": "2026-11-03T17:13:56.266Z",
"elapsedMs": 134
},
{
"stage": "inspect-root",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:13:56.267Z",
"finishedAt": "2026-11-03T17:13:56.297Z",
"elapsedMs": 31
},
{
"stage": "inspect-host",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:13:56.297Z",
"finishedAt": "2026-11-03T17:13:59.827Z",
"elapsedMs": 3530
},
{
"stage": "preparing",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:14:35.216Z",
"finishedAt": "2026-11-03T17:14:35.222Z",
"elapsedMs": 6
},
{
"stage": "verify-candidate",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:14:35.222Z",
"finishedAt": "2026-11-03T17:14:35.252Z",
"elapsedMs": 30
},
{
"stage": "verify-candidate",
"scope": "full-manifest",
"counts": {
"completedFiles": 10949,
"totalFiles": 10949,
"completedBytes": 657252428,
"totalBytes": 657252428,
"currentPath": "scripts/upgrade-running.mjs"
},
"startedAt": "2026-11-03T17:14:35.252Z",
"finishedAt": "2026-11-03T17:17:10.939Z",
"elapsedMs": 155688
},
{
"stage": "verify-candidate",
"scope": "directory-inventory",
"counts": {
"completedFiles": 10950,
"totalFiles": 10950,
"completedBytes": null,
"totalBytes": null,
"currentPath": "THIRD_PARTY_NOTICES.txt"
},
"startedAt": "2026-11-03T17:17:10.939Z",
"finishedAt": "2026-11-03T17:17:51.843Z",
"elapsedMs": 40904
},
{
"stage": "verify-candidate",
"scope": "loaded-inputs",
"counts": {
"completedFiles": 22,
"totalFiles": 22,
"completedBytes": 9321617,
"totalBytes": 9321617,
"currentPath": "dist-server/server.mjs"
},
"startedAt": "2026-11-03T17:17:51.843Z",
"finishedAt": "2026-11-03T17:17:53.084Z",
"elapsedMs": 1241
},
{
"stage": "verify-current",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:17:53.084Z",
"finishedAt": "2026-11-03T17:17:53.121Z",
"elapsedMs": 37
},
{
"stage": "verify-current",
"scope": "full-manifest",
"counts": {
"completedFiles": 10944,
"totalFiles": 10944,
"completedBytes": 657071308,
"totalBytes": 657071308,
"currentPath": "scripts/upgrade-running.mjs"
},
"startedAt": "2026-11-03T17:17:53.121Z",
"finishedAt": "2026-11-03T17:20:29.767Z",
"elapsedMs": 156647
},
{
"stage": "verify-current",
"scope": "directory-inventory",
"counts": {
"completedFiles": 10945,
"totalFiles": 10945,
"completedBytes": null,
"totalBytes": null,
"currentPath": "THIRD_PARTY_NOTICES.txt"
},
"startedAt": "2026-11-03T17:20:29.767Z",
"finishedAt": "2026-11-03T17:21:10.694Z",
"elapsedMs": 40927
},
{
"stage": "verify-current",
"scope": "loaded-inputs",
"counts": {
"completedFiles": 22,
"totalFiles": 22,
"completedBytes": 9306746,
"totalBytes": 9306746,
"currentPath": "dist-server/server.mjs"
},
"startedAt": "2026-11-03T17:21:10.694Z",
"finishedAt": "2026-11-03T17:21:14.556Z",
"elapsedMs": 3862
},
{
"stage": "snapshot-data",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:21:14.556Z",
"finishedAt": "2026-11-03T17:21:37.148Z",
"elapsedMs": 22592
},
{
"stage": "retain-host",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:21:37.148Z",
"finishedAt": "2026-11-03T17:21:41.330Z",
"elapsedMs": 4182
},
{
"stage": "verify-data",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:21:41.330Z",
"finishedAt": "2026-11-03T17:21:48.048Z",
"elapsedMs": 6717
},
{
"stage": "verify-current",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:21:48.048Z",
"finishedAt": "2026-11-03T17:21:48.075Z",
"elapsedMs": 27
},
{
"stage": "verify-current",
"scope": "full-manifest",
"counts": {
"completedFiles": 10944,
"totalFiles": 10944,
"completedBytes": 657071308,
"totalBytes": 657071308,
"currentPath": "scripts/upgrade-running.mjs"
},
"startedAt": "2026-11-03T17:21:48.075Z",
"finishedAt": "2026-11-03T17:22:49.894Z",
"elapsedMs": 61818
},
{
"stage": "verify-current",
"scope": "directory-inventory",
"counts": {
"completedFiles": 10945,
"totalFiles": 10945,
"completedBytes": null,
"totalBytes": null,
"currentPath": "THIRD_PARTY_NOTICES.txt"
},
"startedAt": "2026-11-03T17:22:49.894Z",
"finishedAt": "2026-11-03T17:23:30.779Z",
"elapsedMs": 40885
},
{
"stage": "verify-current",
"scope": "loaded-inputs",
"counts": {
"completedFiles": 22,
"totalFiles": 22,
"completedBytes": 9306746,
"totalBytes": 9306746,
"currentPath": "dist-server/server.mjs"
},
"startedAt": "2026-11-03T17:23:30.779Z",
"finishedAt": "2026-11-03T17:23:30.909Z",
"elapsedMs": 130
},
{
"stage": "verify-candidate",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:23:30.909Z",
"finishedAt": "2026-11-03T17:23:30.932Z",
"elapsedMs": 23
},
{
"stage": "verify-candidate",
"scope": "full-manifest",
"counts": {
"completedFiles": 10949,
"totalFiles": 10949,
"completedBytes": 657252428,
"totalBytes": 657252428,
"currentPath": "scripts/upgrade-running.mjs"
},
"startedAt": "2026-11-03T17:23:30.932Z",
"finishedAt": "2026-11-03T17:25:56.919Z",
"elapsedMs": 145988
},
{
"stage": "verify-candidate",
"scope": "directory-inventory",
"counts": {
"completedFiles": 10950,
"totalFiles": 10950,
"completedBytes": null,
"totalBytes": null,
"currentPath": "THIRD_PARTY_NOTICES.txt"
},
"startedAt": "2026-11-03T17:25:56.919Z",
"finishedAt": "2026-11-03T17:26:37.873Z",
"elapsedMs": 40954
},
{
"stage": "verify-candidate",
"scope": "loaded-inputs",
"counts": {
"completedFiles": 22,
"totalFiles": 22,
"completedBytes": 9321617,
"totalBytes": 9321617,
"currentPath": "dist-server/server.mjs"
},
"startedAt": "2026-11-03T17:26:37.873Z",
"finishedAt": "2026-11-03T17:26:38.012Z",
"elapsedMs": 139
},
{
"stage": "publish-pointer",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:26:38.012Z",
"finishedAt": "2026-11-03T17:26:38.025Z",
"elapsedMs": 14
},
{
"stage": "verify-publication",
"scope": null,
"counts": null,
"startedAt": "2026-11-03T17:26:38.025Z",
"finishedAt": "2026-11-03T17:26:38.073Z",
"elapsedMs": 48
}
],
"process": {
"pid": 15272,
"createdAt": "2026-11-03T17:14:33.493Z"
},
"candidate": {
"root": "C:\\Users\\wangy\\AppData\\Local\\Programs\\AgentWorkbench\\versions\\Agent-Workbench-1.24.7-win-x64-20261103164618190",
"version": "1.24.7",
"manifestSha256": "03683cabd6b88cabdb3eb4cd159ffbabdf1b37e1576354ec3d6433d4ca773be9",
"runtimeIdentity": "212aba55ceb5eb08bb7985dd346bfb3967f5c7a36b2bc848b7e149adb50edf85"
},
"state": "completed",
"stage": "completed",
"scope": null,
"counts": null,
"interruptible": false,
"label": "版本激活完成",
"detail": {
"activeVersion": "1.24.7",
"receiptFile": "C:\\Users\\wangy\\AppData\\Local\\Programs\\AgentWorkbench\\.installation\\operation-a9edbc42-e65b-44f3-bfed-a93e7a8324c4.json"
},
"stageStartedAt": "2026-11-03T17:26:38.073Z",
"updatedAt": "2026-11-03T17:26:38.073Z",
"elapsedMs": 869425
},
"desktopWindows": 0,
"inspectorClosed": true,
"sourceCommit": "5dcd6d7c3f4f465071ae1156bb1e0bf5a28f48e2",
"mode": "install",
"startedAt": "2026-11-03T17:11:57.514Z",
"finishedAt": "2026-11-03T17:26:45.829Z",
"realModelRequests": 0,
"scope": "Public API/full original row/export/artifact hashes; installer retains original running host and Provider. No credential extraction or task replay."
}
@@ -0,0 +1,4 @@
{
"code": 0,
"finishedAt": "2026-11-03T17:26:38.189Z"
}
File diff suppressed because it is too large Load Diff
@@ -0,0 +1,21 @@
{
"schema": 1,
"app": "agent-workbench-installation",
"installId": "74c11bdd-4895-459b-8403-097f79980b41",
"revision": 22,
"dataRoot": "C:\\Users\\wangy\\AppData\\Local\\Packages\\OpenAI.Codex_2p2nqsd0c76g0\\LocalCache\\Local\\AgentWorkbench",
"port": 4318,
"active": {
"directory": "Agent-Workbench-1.24.6-win-x64-20261103151145294",
"version": "1.24.6",
"manifestSha256": "64aa8b2ae90df30eb33b9d2f60b458c3fcf61de7cbc778ac2d08fd0f7437127d",
"runtimeIdentity": "2c662dee28c0c61b57e2f77dbb546667af7b6938aaf848e5a6facde973cad2e1"
},
"previous": {
"directory": "Agent-Workbench-1.24.5-win-x64-20261103133956264",
"version": "1.24.5",
"manifestSha256": "44e442109062fd2f33891aa414ff7b328f951948de1836cd1caf9b775342ec81",
"runtimeIdentity": "ff2a8b8d69648a32740c9725e2d88a4e4c34bd95d69497798fdd67409a45d7d9"
},
"activatedAt": "2026-11-03T15:47:56.270Z"
}
File diff suppressed because one or more lines are too long
@@ -0,0 +1,84 @@
{
"portableRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.24.7-win-x64-20261103164618190",
"sourceRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.24.7-source-20261103164618190",
"evidenceRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.24.7-evidence-20261103164618190",
"portable": {
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.24.7-win-x64-20261103164618190.zip",
"bytes": 242883511,
"sha256": "9f6ec8ae971d9207b34d5a26c2b8e3c09e63b3138793ef25d898ce6e3641b23a"
},
"source": {
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.24.7-source-20261103164618190.zip",
"bytes": 1020471144,
"sha256": "e27ae51497a2a42632486a4b852e98a32c2d0877e9874f85ad1a01b802f49f1e"
},
"evidence": {
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.24.7-evidence-20261103164618190.zip",
"bytes": 779265448,
"sha256": "064f6bd47d6900f63706300ce1594b859ef0a92eebb41f6d656b386ad2981291"
},
"documentation": {
"sourceCommit": "5dcd6d7c3f4f465071ae1156bb1e0bf5a28f48e2",
"originalFiles": 20728,
"originalBytes": 2443917643,
"shippedFiles": 154,
"shippedBytes": 1170702,
"evidenceManifestSha256": "dba6ac06d50083cfcc9724f5bd56a2befd0f1e69eccd3208c58e63632472f360"
},
"installer": {
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.24.7-win-x64-20261103164618190-Setup.exe",
"bytes": 213284660,
"sha256": "a2ea5ce3e249f5aeecc7bb5a8972746a214a59343d95d5828d6a0707a67dc2c2",
"compiler": {
"version": "6.3.1",
"sha256": "3b22d36fc7e7c421d8caf780008ecdb42b14b3f0371b9bafe3cad64e6648d9cf"
},
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\installer-build-2bda8bad-f0af-4582-8304-200b8e7a1078\\installer.iss",
"version": "1.24.7",
"candidateDirectory": "Agent-Workbench-1.24.7-win-x64-20261103164618190",
"qaFixture": null
},
"desktopRuntime": {
"version": "45.2.0",
"sha256": "010ea51bfb528c8143513439aa93fb40754403daa2023f9decd4517da71706ab",
"bytes": 249810432,
"executable": "desktop/AgentWorkbench.exe",
"scope": "Fixed original Windows Electron runtime, full license texts, isolated application loader and native bootstrap; original executable hash unchanged."
},
"nativeLauncher": {
"bytes": 12288,
"sha256": "e3d31cb7ce5009d3ca5a96d04dc16b1add0bd1f5cde7e1304461de6713d37463",
"inputs": {
"compiler": "46809206887326d2d24db1eff1f3064de972c3451abe766b49111450a5e08e00",
"source": "1c8e7d7ee5c383897bec48e9aadb8051b47c09686f74bbf34e0b6b695cd01ae5",
"icon": "f1879074e45f01c480ad9bf50b818637cc60ad56526056c886f808a4812d7b0c"
},
"compiler": "Windows .NET Framework C# compiler; source-bound cached original output, no deterministic-build claim."
},
"nodeVersion": "v24.22.1",
"reaRuntime": {
"version": "6.8.0",
"protocolVersion": "2025-03-26",
"packages": 149,
"files": 8413,
"bytes": 142395402,
"lockSha256": "687fb3fa3089a12f239dd5e87d9fa07df55a87df4935931d495a92600a1b0f6e"
},
"portableFiles": 10949,
"licenseTexts": 135,
"browserRuntime": {
"playwrightVersion": "1.64.1",
"packages": [
{
"name": "playwright",
"version": "1.64.1",
"files": 60
},
{
"name": "playwright-core",
"version": "1.64.1",
"files": 115
}
]
}
}
@@ -0,0 +1 @@
{"passed":true,"cases":["1440: actual reported overrun persists, zero unknown use stays distinct, and preview does not dispatch or replace the original Task","390: actual reported overrun persists, zero unknown use stays distinct, and preview does not dispatch or replace the original Task"],"screenshots":["D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-budget-ui-25db6351-c80f-4abb-9286-d10b0c3c1ff0\\reported-overrun-1440.png","D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-budget-ui-25db6351-c80f-4abb-9286-d10b0c3c1ff0\\reported-overrun-390.png"],"errors":[],"realModelRequests":0,"ownedHttpRequests":1,"modelObservation":{"schemaVersion":1,"source":"task_run_consumption","taskId":"ca25aebb-ed6a-47c9-b6ff-3950098b5fe0","status":"observed","identityScope":"request_routes_and_upstream_reported_identifiers_not_model_weights","responseIdentity":{"counts":{"incomplete":1},"reportedModels":[],"allAccepted":false},"model":"owned-budget-ui","models":["owned-budget-ui"],"observedAttempts":1,"recordedRequests":1,"routes":[{"connectionId":"be582245-7e15-4a95-9b72-ed89968e55b1","model":"owned-budget-ui","protocol":"chat-completions","requests":1,"reportedUsageRequests":1,"unknownUsageRequests":0}],"attempts":[{"runId":"9731b662-21b0-43ed-bb4a-e80db8f43010","attemptId":"4612cc2c-efb3-4346-b2ba-60ef728e9bf9","requestId":"503cded9-5368-4783-b503-517356153758","model":"owned-budget-ui","protocol":"chat-completions","connectionId":"be582245-7e15-4a95-9b72-ed89968e55b1","outcome":"failed","requestedModel":"owned-budget-ui","modelIdentity":{"schemaVersion":1,"requestedModel":"owned-budget-ui","reportedModels":[],"status":"incomplete","complete":false,"invalidIdentifiers":false,"policy":"record","scope":"Identifiers in the outgoing request and upstream response; neither equal identifiers nor a declared mapping authenticate model weights."},"usageStatus":"reported","usage":{"prompt_tokens":55000,"completion_tokens":17,"total_tokens":55017}}],"issues":[]},"scope":"Actual compiled host and owned HTTP reported overrun, persistence and preview UI; real-model autonomy/Purge benefit remain separate.","directory":"D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-budget-ui-25db6351-c80f-4abb-9286-d10b0c3c1ff0"}
@@ -0,0 +1,45 @@
PASS desktop request estimate, actual usage, complete goal and history identity are visible
PASS paged complete history opens actual tool result artifacts without executing tools
PASS historical and resumed Runs retain their own request records
PASS 390px history, original JSON and sent-body dialogs fit and retain close controls
PASS late old history failure stays in its original scope after switching projects
PASS late old Run failure stays in its original scope after switching projects
PASS late old request-body failure stays in its original scope after switching projects
PASS context-limit pause explains the intact requirement and explicit continuation
PASS 390px repeated-read pause explains the cause and explicit continuation retains the same Task
PASS desktop working evidence shows actual code, original result provenance and the current Task
PASS 390px working evidence and sent request stay readable and closing the artifact retains task navigation
{
"success": true,
"version": "1.24.7",
"testedAt": "2026-11-03T16:44:40.801Z",
"passed": [
"desktop request estimate, actual usage, complete goal and history identity are visible",
"paged complete history opens actual tool result artifacts without executing tools",
"historical and resumed Runs retain their own request records",
"390px history, original JSON and sent-body dialogs fit and retain close controls",
"late old history failure stays in its original scope after switching projects",
"late old Run failure stays in its original scope after switching projects",
"late old request-body failure stays in its original scope after switching projects",
"context-limit pause explains the intact requirement and explicit continuation",
"390px repeated-read pause explains the cause and explicit continuation retains the same Task",
"desktop working evidence shows actual code, original result provenance and the current Task",
"390px working evidence and sent request stay readable and closing the artifact retains task navigation"
],
"pageErrors": [],
"requestFailures": [],
"screenshots": [
".qa/screenshots/reading-s-desktop-plan.png",
".qa/screenshots/reading-s-desktop-history.png",
".qa/screenshots/reading-s-mobile-history.png",
".qa/screenshots/reading-s-mobile-original.png",
".qa/screenshots/reading-s-mobile-request-body.png",
".qa/screenshots/reading-s-mobile-read-pause.png",
".qa/screenshots/reading-s-desktop-working-evidence.png",
".qa/screenshots/reading-s-mobile-working-evidence.png"
],
"root": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\context-ui-008294dc-c99b-4373-8c4a-0c45eda7bee6",
"data": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\context-ui-008294dc-c99b-4373-8c4a-0c45eda7bee6\\data",
"mode": "saved synthetic context evidence; no real model or project commands executed by this UI verifier",
"browser": "156.0.4314.50"
}
@@ -0,0 +1,8 @@
PASS sidebar and header share the authoritative multi-model selector; selection persists without a request
PASS add from primary entry keeps the previous default, supports a new selection and survives reload without paid probes
PASS budget-only saves preserve the pool and original connection editing is explicitly secondary and cancellable
PASS actual pool request reaches only the selected endpoint; changing defaults preserves the sealed old Run and shows both identities
PASS opening refreshes another tab selection and choosing a real connection exits demo without losing profiles
PASS 390px header and sidebar both add and switch with long model labels and no page overflow
PASS active request locks route and add controls and server rejects budget overrides across sessions
{"success":true,"directory":"D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-selection-k-d11c8059-7775-4e7b-b168-e7ba5a123ae1","checks":7}
Binary file not shown.

After

Width:  |  Height:  |  Size: 129 KiB

@@ -0,0 +1,388 @@
Window: "Agent Workbench · 本地编程工作台", App: AgentWorkbench.exe.
0 窗口 Agent Workbench · 本地编程工作台 Secondary Actions: Raise
1 窗格 (disabled)
2 区域 Agent Workbench · 本地编程工作台
3 区域
4 区域
5 区域
6 区域
7 菜单栏
8 按钮 工作台
9 按钮 编辑 ID: view_1
10 区域
11 区域
13 文档 Agent Workbench · 本地编程工作台 Value: http://127.0.0.1:4318/?desktop=1&release=1.24.7 ID: RootWebArea
14 组 ID: root
15 补充
16 文本 AW
17 文本 Agent Workbench
18 文本 本地编程工作台 · 正式版
19 按钮 收起项目栏 Description: 收起项目栏
20 文本 当前项目
21 组合框 (settable, string) 当前项目 Value: starter-project
22 按钮 打开项目 Description: 打开项目
23 文本 D:\Ai-project\REA逆向\Agent-Workbench\releases\Agent-Workbench-0.1.0-win-x64\examples\starter-project
24 文本 任务会话
25 按钮 新建任务 Description: 新建任务
26 按钮 新建任务 ↵
27 按钮 创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D
28 按钮 归档会话 创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D Description: 归档会话 创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D
29 按钮 了解这个项目,并提出一项改进
30 按钮 归档会话 了解这个项目,并提出一项改进 Description: 归档会话 了解这个项目,并提出一项改进
31 文本 项目文件
32 文本 5
33 按钮 刷新文件 Description: 刷新文件
34 编辑 (settable, string) 筛选文件 Description: 筛选文件
35 文本 ⌃ K
90 按钮 src Description: src
91 按钮 catalog.js Description: src/catalog.js
92 按钮 tests Description: tests
93 按钮 catalog.test.mjs Description: tests/catalog.test.mjs
94 按钮 package.json Description: package.json
95 按钮 pelican-cycle.html Description: pelican-cycle.html
96 按钮 README.md Description: README.md
36 按钮 机制研究与设计
37 按钮 项目记忆
38 按钮 运行诊断
39 按钮 选择与管理模型
40 文本 本地服务已连接
41 按钮 版本与更新
42 主要
43 章节标题
44 文本 starter-project
45 文本 任务
46 标题 创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D
47 按钮 手机连接
48 按钮 桌面宿主
49 按钮 打开模型中心
50 按钮 切换当前模型 Description: 选择模型、模型池或添加连接
51 文本 本地项目
97 文本 已完成
53 按钮 运行诊断 Description: 运行诊断
54 按钮 项目技能 Description: 项目技能
55 按钮 项目MCP工具 Description: 项目MCP工具
56 按钮 查看任务记录 Description: 查看任务记录
57 按钮 隔离协作 Description: 隔离协作
58 按钮 工程评测 Description: 工程评测
59 按钮 本地代理引擎 Description: 本地代理引擎
60 按钮 REA 工件研究 Description: REA 工件研究
98 按钮 导出会话 Description: 导出会话
61 区域 任务对话
99 文本 你
100 文本 你的任务
101 time
102 文本 20:08
103 文本 创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D动画,无需测试。
104 按钮 模型响应已收到 模型服务 5.6s
105 文本 AW
106 文本 编程助手
107 time
108 文本 20:09
109 文本 我会制作一个可直接在浏览器打开的独立 HTML,用 SVG 绘制鹈鹕、自行车和沿途风景,加入踩踏、车轮转动与背景移动动画。按你的要求,不运行测试。
110 按钮 识别工程与Git基线 0.1s
111 按钮 模型连接失败 模型服务 · 第 2 次 60.0s
112 组 模型输出片段
113 文本 输出未完成
114 文本 Chat Completions
115 文本 ·
116 文本 0
117 文本 字符
118 文本 · 1 个工具参数片段
119 文本 片段仅供查看。完整响应核对前不执行工具;中断后保留原记录。
120 文本 ·
121 文本 工作台
122 time
123 文本 20:11
124 警报
125 文本 模型连接
126 文本 模型服务响应超时
127 文本 模型服务响应超时。项目文件和已完成的工具记录已保留。 已接收的流片段保留,请求不会自动重发。
128 文本 可以重试上次模型请求;若持续失败,在模型连接中检查服务地址和连接状态。
129 文本 发生在请求模型响应时
130 文本 1
131 文本 次请求
132 详细信息
133 开合三角标记 查看错误详情
134 文本 查看错误详情
135 按钮 查看运行诊断
136 文本 ·
137 文本 工作台
138 time
139 文本 20:47
140 文本 从已保存的执行记录继续,已完成的工具结果保留。
141 按钮 模型连接失败 模型服务 · 第 2 次 60.0s
142 组 模型输出片段
143 文本 输出未完成
144 文本 Chat Completions
145 文本 ·
146 文本 0
147 文本 字符
148 文本 · 1 个工具参数片段
149 文本 片段仅供查看。完整响应核对前不执行工具;中断后保留原记录。
150 文本 ·
151 文本 工作台
152 time
153 文本 20:49
154 警报
155 文本 模型连接
156 文本 模型服务响应超时
157 文本 模型服务响应超时。项目文件和已完成的工具记录已保留。 已接收的流片段保留,请求不会自动重发。
158 文本 可以重试上次模型请求;若持续失败,在模型连接中检查服务地址和连接状态。
159 文本 发生在请求模型响应时
160 文本 1
161 文本 次请求
162 详细信息
163 开合三角标记 查看错误详情
164 文本 查看错误详情
165 按钮 查看运行诊断
166 文本 ·
167 文本 工作台
168 time
169 文本 21:03
170 文本 从已保存的执行记录继续,已完成的工具结果保留。
171 按钮 模型连接失败 模型服务 · 第 2 次 60.0s
172 组 模型输出片段
173 文本 输出未完成
174 文本 Chat Completions
175 文本 ·
176 文本 0
177 文本 字符
178 文本 · 1 个工具参数片段
179 文本 片段仅供查看。完整响应核对前不执行工具;中断后保留原记录。
180 文本 ·
181 文本 工作台
182 time
183 文本 21:05
184 警报
185 文本 模型连接
186 文本 模型服务响应超时
187 文本 模型服务响应超时。项目文件和已完成的工具记录已保留。 已接收的流片段保留,请求不会自动重发。
188 文本 可以重试上次模型请求;若持续失败,在模型连接中检查服务地址和连接状态。
189 文本 发生在请求模型响应时
190 文本 1
191 文本 次请求
192 详细信息
193 开合三角标记 查看错误详情
194 文本 查看错误详情
195 按钮 查看运行诊断
196 文本 ·
197 文本 工作台
198 time
199 文本 21:26
200 文本 从已保存的执行记录继续,已完成的工具结果保留。
201 按钮 模型响应已收到 模型服务 121.1s
202 按钮 准备文件改动 pelican-cycle.html 0.0s
203 文本 整理请求上下文
204 文本 完整请求输入估算 13568 → 8741 tokens;省略 1 个完整消息组,原历史和参数已保存为SHA256工件。估算与服务实际用量分别记录。
205 按钮 模型响应已收到 模型服务 8.6s
206 文本 AW
207 文本 编程助手
208 time
209 文本 21:28
210 文本 已准备好
211 文本 pelican-cycle.html
212 文本 的文件提案,内容为 SVG 绘制的鹈鹕骑自行车 2D 动画。
213 文本 请在改动面板查看差异并点击“应用”;应用后即可用浏览器打开该 HTML。当前提案尚未写入项目。
214 文本 按你的要求,未运行测试。
215 文本 ·
216 文本 工作台
217 time
218 文本 21:28
219 文本 文件已审查并应用:pelican-cycle.html。按照原要求未运行测试,可在文件列表打开结果。
220 文本 已应用 pelican-cycle.html
221 文本 写入前基线检查通过,已创建新文件。
69 章节页脚
70 按钮 添加截图
71 文本 每任务1张 · PNG/JPEG · 原图≤2MB
72 编辑 (settable, string) 任务内容 Description: 描述项目任务,或询问工作台的执行状态与报错…
73 文本 starter-project
74 文本 高性能
75 按钮 (disabled) 发送任务
76 文本 Ctrl + Enter 发送 · 会话自动保存
77 文本 上下文估算 8,741 tokens
78 补充
79 缩略图 调整检视区宽度 Value: 50
80 按钮 文件
81 按钮 改动
82 按钮 轨迹
83 按钮 任务
84 按钮 索引
85 按钮 浏览
86 按钮 放大检视区 Description: 放大检视区
87 标题 查看项目文件
88 文本 从左侧打开文件,或让助手先读取项目。
89 文本 读取范围限定在当前项目
12 区域
Document text: ```
AW
Agent Workbench
本地编程工作台 · 正式版
收起项目栏
当前项目
当前项目
打开项目
D:\Ai-project\REA逆向\Agent-Workbench\releases\Agent-Workbench-0.1.0-win-x64\examples\starter-projectD:\Ai-project\REA逆向\Agent-Wor…
任务会话
新建任务
新建任务
↵
创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D创建一个HTML,内容是SVG…
归档会话 创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D
了解这个项目,并提出一项改进了解这个项目,并提出一项…
归档会话 了解这个项目,并提出一项改进
项目文件 5
刷新文件
筛选文件
⌃ K
src
catalog.js
tests
catalog.test.mjs
package.json
pelican-cycle.html
README.md
机制研究与设计
项目记忆
运行诊断
模型与连接
高性能
本地服务已连接
v1.24.7
starter-project任务
创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D创建一个HTML,内容是SVG绘制一…
手机连接
桌面宿主
模型中心
高性能
本地项目
已完成
运行诊断
项目技能
项目MCP工具
查看任务记录
隔离协作
工程评测
本地代理引擎
REA 工件研究
导出会话
你
你的任务
20:08
创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D动画,无需测试。
模型响应已收到
模型服务
5.6s
AW
编程助手
20:09
我会制作一个可直接在浏览器打开的独立 HTML,用 SVG 绘制鹈鹕、自行车和沿途风景,加入踩踏、车轮转动与背景移动动画。按你的要求,不运行测试。
识别工程与Git基线
0.1s
模型连接失败
模型服务 · 第 2 次
60.0s
输出未完成
Chat Completions · 0 字符 · 1 个工具参数片段
片段仅供查看。完整响应核对前不执行工具;中断后保留原记录。
·
工作台
20:11
模型连接
模型服务响应超时
模型服务响应超时。项目文件和已完成的工具记录已保留。 已接收的流片段保留,请求不会自动重发。
可以重试上次模型请求;若持续失败,在模型连接中检查服务地址和连接状态。
发生在请求模型响应时
1 次请求
查看错误详情
查看运行诊断
·
工作台
20:47
从已保存的执行记录继续,已完成的工具结果保留。
模型连接失败
模型服务 · 第 2 次
60.0s
输出未完成
Chat Completions · 0 字符 · 1 个工具参数片段
片段仅供查看。完整响应核对前不执行工具;中断后保留原记录。
·
工作台
20:49
模型连接
模型服务响应超时
模型服务响应超时。项目文件和已完成的工具记录已保留。 已接收的流片段保留,请求不会自动重发。
可以重试上次模型请求;若持续失败,在模型连接中检查服务地址和连接状态。
发生在请求模型响应时
1 次请求
查看错误详情
查看运行诊断
·
工作台
21:03
从已保存的执行记录继续,已完成的工具结果保留。
模型连接失败
模型服务 · 第 2 次
60.0s
输出未完成
Chat Completions · 0 字符 · 1 个工具参数片段
片段仅供查看。完整响应核对前不执行工具;中断后保留原记录。
·
工作台
21:05
模型连接
模型服务响应超时
模型服务响应超时。项目文件和已完成的工具记录已保留。 已接收的流片段保留,请求不会自动重发。
可以重试上次模型请求;若持续失败,在模型连接中检查服务地址和连接状态。
发生在请求模型响应时
1 次请求
查看错误详情
查看运行诊断
·
工作台
21:26
从已保存的执行记录继续,已完成的工具结果保留。
模型响应已收到
模型服务
121.1s
准备文件改动
pelican-cycle.html
0.0s
整理请求上下文
完整请求输入估算 13568 → 8741 tokens;省略 1 个完整消息组,原历史和参数已保存为SHA256工件。估算与服务实际用量分别记录。
模型响应已收到
模型服务
8.6s
AW
编程助手
21:28
已准备好 pelican-cycle.html 的文件提案,内容为 SVG 绘制的鹈鹕骑自行车 2D 动画。
请在改动面板查看差异并点击“应用”;应用后即可用浏览器打开该 HTML。当前提案尚未写入项目。
按你的要求,未运行测试。
·
工作台
21:28
文件已审查并应用:pelican-cycle.html。按照原要求未运行测试,可在文件列表打开结果。
已应用 pelican-cycle.html
写入前基线检查通过,已创建新文件。
添加截图
每任务1张 · PNG/JPEG · 原图≤2MB
任务内容
starter-project
高性能
发送任务
Ctrl + Enter 发送 · 会话自动保存
上下文估算 8,741 tokens
调整检视区宽度
文件
改动
轨迹
任务
索引
浏览
放大检视区

查看项目文件
从左侧打开文件,或让助手先读取项目。
读取范围限定在当前项目
```
Binary file not shown.

After

Width:  |  Height:  |  Size: 108 KiB

@@ -0,0 +1,484 @@
Window: "Agent Workbench · 本地编程工作台", App: AgentWorkbench.exe.
0 窗口 Agent Workbench · 本地编程工作台 Secondary Actions: Raise
1 窗格 (disabled)
2 区域 Agent Workbench · 本地编程工作台
3 区域
4 区域
5 区域
6 区域
7 菜单栏
8 按钮 工作台
9 按钮 编辑 ID: view_1
10 区域
11 区域
13 文档 Agent Workbench · 本地编程工作台 Value: http://127.0.0.1:4318/?desktop=1&release=1.24.6 ID: RootWebArea
225 组
14 组 ID: root
15 补充
16 文本 AW
17 文本 Agent Workbench
18 文本 本地编程工作台 · 正式版
19 按钮 收起项目栏 Description: 收起项目栏
20 文本 当前项目
21 组合框 (settable, string) 当前项目 Value: starter-project
22 按钮 打开项目 Description: 打开项目
88 文本 D:\Ai-project\REA逆向\Agent-Workbench\releases\Agent-Workbench-0.1.0-win-x64\examples\starter-project
23 文本 任务会话
24 按钮 新建任务 Description: 新建任务
25 按钮 新建任务 ↵
89 按钮 创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D
90 按钮 归档会话 创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D Description: 归档会话 创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D
91 按钮 了解这个项目,并提出一项改进
92 按钮 归档会话 了解这个项目,并提出一项改进 Description: 归档会话 了解这个项目,并提出一项改进
27 文本 项目文件
28 文本 5
29 按钮 刷新文件 Description: 刷新文件
30 编辑 (settable, string) 筛选文件 Description: 筛选文件
31 文本 ⌃ K
93 按钮 src Description: src
94 按钮 catalog.js Description: src/catalog.js
95 按钮 tests Description: tests
96 按钮 catalog.test.mjs Description: tests/catalog.test.mjs
97 按钮 package.json Description: package.json
98 按钮 pelican-cycle.html Description: pelican-cycle.html
99 按钮 README.md Description: README.md
32 按钮 机制研究与设计
33 按钮 项目记忆
34 按钮 运行诊断
35 按钮 选择与管理模型
36 文本 本地服务已连接
37 按钮 版本与更新
38 主要
39 章节标题
40 文本 starter-project
41 文本 任务
42 标题 创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D
43 按钮 手机连接
44 按钮 桌面宿主
45 按钮 打开模型中心
46 按钮 切换当前模型 Description: 选择模型、模型池或添加连接
47 文本 本地项目
100 文本 已完成
49 按钮 运行诊断 Description: 运行诊断
50 按钮 项目技能 Description: 项目技能
51 按钮 项目MCP工具 Description: 项目MCP工具
52 按钮 查看任务记录 Description: 查看任务记录
53 按钮 隔离协作 Description: 隔离协作
54 按钮 工程评测 Description: 工程评测
55 按钮 本地代理引擎 Description: 本地代理引擎
56 按钮 REA 工件研究 Description: REA 工件研究
101 按钮 导出会话 Description: 导出会话
57 区域 任务对话
102 文本 你
103 文本 你的任务
104 time
105 文本 20:08
106 文本 创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D动画,无需测试。
107 按钮 模型响应已收到 模型服务 5.6s
108 文本 AW
109 文本 编程助手
110 time
111 文本 20:09
112 文本 我会制作一个可直接在浏览器打开的独立 HTML,用 SVG 绘制鹈鹕、自行车和沿途风景,加入踩踏、车轮转动与背景移动动画。按你的要求,不运行测试。
113 按钮 识别工程与Git基线 0.1s
114 按钮 模型连接失败 模型服务 · 第 2 次 60.0s
115 组 模型输出片段
116 文本 输出未完成
117 文本 Chat Completions
118 文本 ·
119 文本 0
120 文本 字符
121 文本 · 1 个工具参数片段
122 文本 片段仅供查看。完整响应核对前不执行工具;中断后保留原记录。
123 文本 ·
124 文本 工作台
125 time
126 文本 20:11
127 警报
128 文本 模型连接
129 文本 模型服务响应超时
130 文本 模型服务响应超时。项目文件和已完成的工具记录已保留。 已接收的流片段保留,请求不会自动重发。
131 文本 可以重试上次模型请求;若持续失败,在模型连接中检查服务地址和连接状态。
132 文本 发生在请求模型响应时
133 文本 1
134 文本 次请求
135 详细信息
136 开合三角标记 查看错误详情
137 文本 查看错误详情
138 按钮 查看运行诊断
139 文本 ·
140 文本 工作台
141 time
142 文本 20:47
143 文本 从已保存的执行记录继续,已完成的工具结果保留。
144 按钮 模型连接失败 模型服务 · 第 2 次 60.0s
145 组 模型输出片段
146 文本 输出未完成
147 文本 Chat Completions
148 文本 ·
149 文本 0
150 文本 字符
151 文本 · 1 个工具参数片段
152 文本 片段仅供查看。完整响应核对前不执行工具;中断后保留原记录。
153 文本 ·
154 文本 工作台
155 time
156 文本 20:49
157 警报
158 文本 模型连接
159 文本 模型服务响应超时
160 文本 模型服务响应超时。项目文件和已完成的工具记录已保留。 已接收的流片段保留,请求不会自动重发。
161 文本 可以重试上次模型请求;若持续失败,在模型连接中检查服务地址和连接状态。
162 文本 发生在请求模型响应时
163 文本 1
164 文本 次请求
165 详细信息
166 开合三角标记 查看错误详情
167 文本 查看错误详情
168 按钮 查看运行诊断
169 文本 ·
170 文本 工作台
171 time
172 文本 21:03
173 文本 从已保存的执行记录继续,已完成的工具结果保留。
174 按钮 模型连接失败 模型服务 · 第 2 次 60.0s
175 组 模型输出片段
176 文本 输出未完成
177 文本 Chat Completions
178 文本 ·
179 文本 0
180 文本 字符
181 文本 · 1 个工具参数片段
182 文本 片段仅供查看。完整响应核对前不执行工具;中断后保留原记录。
183 文本 ·
184 文本 工作台
185 time
186 文本 21:05
187 警报
188 文本 模型连接
189 文本 模型服务响应超时
190 文本 模型服务响应超时。项目文件和已完成的工具记录已保留。 已接收的流片段保留,请求不会自动重发。
191 文本 可以重试上次模型请求;若持续失败,在模型连接中检查服务地址和连接状态。
192 文本 发生在请求模型响应时
193 文本 1
194 文本 次请求
195 详细信息
196 开合三角标记 查看错误详情
197 文本 查看错误详情
198 按钮 查看运行诊断
199 文本 ·
200 文本 工作台
201 time
202 文本 21:26
203 文本 从已保存的执行记录继续,已完成的工具结果保留。
204 按钮 模型响应已收到 模型服务 121.1s
205 按钮 准备文件改动 pelican-cycle.html 0.0s
206 文本 整理请求上下文
207 文本 完整请求输入估算 13568 → 8741 tokens;省略 1 个完整消息组,原历史和参数已保存为SHA256工件。估算与服务实际用量分别记录。
208 按钮 模型响应已收到 模型服务 8.6s
209 文本 AW
210 文本 编程助手
211 time
212 文本 21:28
213 文本 已准备好
214 文本 pelican-cycle.html
215 文本 的文件提案,内容为 SVG 绘制的鹈鹕骑自行车 2D 动画。
216 文本 请在改动面板查看差异并点击“应用”;应用后即可用浏览器打开该 HTML。当前提案尚未写入项目。
217 文本 按你的要求,未运行测试。
218 文本 ·
219 文本 工作台
220 time
221 文本 21:28
222 文本 文件已审查并应用:pelican-cycle.html。按照原要求未运行测试,可在文件列表打开结果。
223 文本 已应用 pelican-cycle.html
224 文本 写入前基线检查通过,已创建新文件。
67 章节页脚
68 按钮 添加截图
69 文本 每任务1张 · PNG/JPEG · 原图≤2MB
70 编辑 (settable, string) 任务内容 Description: 描述项目任务,或询问工作台的执行状态与报错…
71 文本 starter-project
72 文本 高性能
73 按钮 (disabled) 发送任务
74 文本 Ctrl + Enter 发送 · 会话自动保存
75 文本 上下文估算 8,741 tokens
76 补充
77 缩略图 调整检视区宽度 Value: 50
78 按钮 文件
79 按钮 改动
80 按钮 轨迹
81 按钮 任务
82 按钮 索引
83 按钮 浏览
84 按钮 放大检视区 Description: 放大检视区
85 标题 查看项目文件
86 文本 从左侧打开文件,或让助手先读取项目。
87 文本 读取范围限定在当前项目
226 对话框 模型中心 Secondary Actions: Raise
227 标题 模型中心
228 文本 连接、模型池与执行健康统一管理。
229 按钮 关闭窗口 Description: 关闭窗口
230 区域 模型控制台
231 文本 MODEL CONTROL
232 标题 选择模型,继续开发。
233 文本 连接、模型池和实际请求使用同一个入口。
234 文本 新任务默认
235 文本 高性能
236 文本 0
237 文本 等待
238 导航 模型中心分类
239 按钮 模型连接
240 按钮 连接诊断
241 按钮 模型池
242 按钮 运行预算
243 按钮 等待时限
244 按钮 请求账本
245 按钮 配置导入
246 按钮 凭据保存
247 按钮 运行巡检
248 按钮 刷新模型中心
250 区域 默认模型选择
251 标题 新任务默认模型
252 文本 选择已保存的连接,或使用模型池调度。切换本身不请求模型。
253 按钮 添加模型
254 文本 当前选择
255 组合框 (settable, string) 新任务默认模型 Value: 高性能 · 2 个连接
256 文本 模型池按成员配置调度;每次实际连接和返回型号保存在请求账本。
257 文本 当前会话最近运行
258 文本 jojocode · gpt-6.1-sol;apikey · claude-opus-5-5
259 文本 这是该次运行封存的请求配置;实际返回型号请看请求账本。更改默认选择不会改写旧记录。
452 区域 模型连接诊断
453 标题 连接诊断
454 文本 检查端点的 DNS、TCP 和 TLS。无需 Key,不发送模型请求。
455 按钮 刷新连接诊断
456 文本 诊断连接
457 组合框 (settable, string) 诊断连接 Value: apikey
458 按钮 开始直连检查
459 文本 https://api.apikey.fan/v1
460 文本
461 文本 · 配置修订
462 文本 11
463 文本 本次只检查端点直连路径,保留证书校验;握手通过不能证明 Key、型号或模型能力可用。检查最多约 12 秒,离开后结果保留。
464 组 本次连接诊断结果
465 标题 先检查连接,再判断问题阶段
466 文本 选择已保存的连接并开始检查。这里会保留解析地址、连接耗时和证书核对结果,原任务不会被重新执行。
467 补充 连接诊断历史
468 标题 近期检查
469 文本 保存最近
470 文本 30
471 文本 条;点击查看原记录。
472 文本 这条连接还没有检查记录。
12 区域
Document text: ```
AW
Agent Workbench
本地编程工作台 · 正式版
收起项目栏
当前项目
当前项目
打开项目
D:\Ai-project\REA逆向\Agent-Workbench\releases\Agent-Workbench-0.1.0-win-x64\examples\starter-projectD:\Ai-project\REA逆向\Agent-Wor…
任务会话
新建任务
新建任务
↵
创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D创建一个HTML,内容是SVG…
归档会话 创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D
了解这个项目,并提出一项改进了解这个项目,并提出一项…
归档会话 了解这个项目,并提出一项改进
项目文件 5
刷新文件
筛选文件
⌃ K
src
catalog.js
tests
catalog.test.mjs
package.json
pelican-cycle.html
README.md
机制研究与设计
项目记忆
运行诊断
模型与连接
高性能
本地服务已连接
v1.24.6
starter-project任务
创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D创建一个HTML,内容是SVG绘制一…
手机连接
桌面宿主
模型中心
高性能
本地项目
已完成
运行诊断
项目技能
项目MCP工具
查看任务记录
隔离协作
工程评测
本地代理引擎
REA 工件研究
导出会话
你
你的任务
20:08
创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D动画,无需测试。
模型响应已收到
模型服务
5.6s
AW
编程助手
20:09
我会制作一个可直接在浏览器打开的独立 HTML,用 SVG 绘制鹈鹕、自行车和沿途风景,加入踩踏、车轮转动与背景移动动画。按你的要求,不运行测试。
识别工程与Git基线
0.1s
模型连接失败
模型服务 · 第 2 次
60.0s
输出未完成
Chat Completions · 0 字符 · 1 个工具参数片段
片段仅供查看。完整响应核对前不执行工具;中断后保留原记录。
·
工作台
20:11
模型连接
模型服务响应超时
模型服务响应超时。项目文件和已完成的工具记录已保留。 已接收的流片段保留,请求不会自动重发。
可以重试上次模型请求;若持续失败,在模型连接中检查服务地址和连接状态。
发生在请求模型响应时
1 次请求
查看错误详情
查看运行诊断
·
工作台
20:47
从已保存的执行记录继续,已完成的工具结果保留。
模型连接失败
模型服务 · 第 2 次
60.0s
输出未完成
Chat Completions · 0 字符 · 1 个工具参数片段
片段仅供查看。完整响应核对前不执行工具;中断后保留原记录。
·
工作台
20:49
模型连接
模型服务响应超时
模型服务响应超时。项目文件和已完成的工具记录已保留。 已接收的流片段保留,请求不会自动重发。
可以重试上次模型请求;若持续失败,在模型连接中检查服务地址和连接状态。
发生在请求模型响应时
1 次请求
查看错误详情
查看运行诊断
·
工作台
21:03
从已保存的执行记录继续,已完成的工具结果保留。
模型连接失败
模型服务 · 第 2 次
60.0s
输出未完成
Chat Completions · 0 字符 · 1 个工具参数片段
片段仅供查看。完整响应核对前不执行工具;中断后保留原记录。
·
工作台
21:05
模型连接
模型服务响应超时
模型服务响应超时。项目文件和已完成的工具记录已保留。 已接收的流片段保留,请求不会自动重发。
可以重试上次模型请求;若持续失败,在模型连接中检查服务地址和连接状态。
发生在请求模型响应时
1 次请求
查看错误详情
查看运行诊断
·
工作台
21:26
从已保存的执行记录继续,已完成的工具结果保留。
模型响应已收到
模型服务
121.1s
准备文件改动
pelican-cycle.html
0.0s
整理请求上下文
完整请求输入估算 13568 → 8741 tokens;省略 1 个完整消息组,原历史和参数已保存为SHA256工件。估算与服务实际用量分别记录。
模型响应已收到
模型服务
8.6s
AW
编程助手
21:28
已准备好 pelican-cycle.html 的文件提案,内容为 SVG 绘制的鹈鹕骑自行车 2D 动画。
请在改动面板查看差异并点击“应用”;应用后即可用浏览器打开该 HTML。当前提案尚未写入项目。
按你的要求,未运行测试。
·
工作台
21:28
文件已审查并应用:pelican-cycle.html。按照原要求未运行测试,可在文件列表打开结果。
已应用 pelican-cycle.html
写入前基线检查通过,已创建新文件。
添加截图
每任务1张 · PNG/JPEG · 原图≤2MB
任务内容
starter-project
高性能
发送任务
Ctrl + Enter 发送 · 会话自动保存
上下文估算 8,741 tokens
调整检视区宽度
文件
改动
轨迹
任务
索引
浏览
放大检视区

查看项目文件
从左侧打开文件,或让助手先读取项目。
读取范围限定在当前项目
模型中心
连接、模型池与执行健康统一管理。
关闭窗口
MODEL CONTROL
选择模型,继续开发。
连接、模型池和实际请求使用同一个入口。
新任务默认
高性能
0 等待
模型连接
连接诊断
模型池
运行预算
等待时限
请求账本
配置导入
凭据保存
运行巡检
刷新模型中心
新任务默认模型
选择已保存的连接,或使用模型池调度。切换本身不请求模型。
添加模型
当前选择
新任务默认模型
模型池按成员配置调度;每次实际连接和返回型号保存在请求账本。
当前会话最近运行
jojocode · gpt-6.1-sol;apikey · claude-opus-5-5
这是该次运行封存的请求配置;实际返回型号请看请求账本。更改默认选择不会改写旧记录。
连接诊断
检查端点的 DNS、TCP 和 TLS。无需 Key,不发送模型请求。
刷新
诊断连接
诊断连接
开始直连检查
https://api.apikey.fan/v1 · 配置修订 11
本次只检查端点直连路径,保留证书校验;握手通过不能证明 Key、型号或模型能力可用。检查最多约 12 秒,离开后结果保留。
先检查连接,再判断问题阶段
选择已保存的连接并开始检查。这里会保留解析地址、连接耗时和证书核对结果,原任务不会被重新执行。
近期检查
保存最近 30 条;点击查看原记录。
这条连接还没有检查记录。
```
@@ -0,0 +1,9 @@
{
"originalExportName": "before.png",
"actualFormat": "JPEG",
"correctedName": "before.jpg",
"bytes": 110051,
"sha256": "df5a7d52590bdc7fe7a97cda6586691e42638d8aa88af899171e5359209858da",
"pixelConversion": false,
"note": "Corrected export extension from actual screenshot bytes before sealing; no recapture or original byte change."
}
Binary file not shown.

After

Width:  |  Height:  |  Size: 126 KiB

@@ -0,0 +1,740 @@
Window: "Agent Workbench · 本地编程工作台", App: AgentWorkbench.exe.
0 窗口 Agent Workbench · 本地编程工作台 Secondary Actions: Raise
1 窗格 (disabled)
2 区域 Agent Workbench · 本地编程工作台
3 区域
4 区域
5 区域
6 区域
7 菜单栏
8 按钮 工作台
9 按钮 编辑 ID: view_1
10 区域
11 区域
13 文档 Agent Workbench · 本地编程工作台 Value: http://127.0.0.1:4318/?desktop=1&release=1.24.7 ID: RootWebArea
222 组
14 组 ID: root
15 补充
16 文本 AW
17 文本 Agent Workbench
18 文本 本地编程工作台 · 正式版
19 按钮 收起项目栏 Description: 收起项目栏
20 文本 当前项目
21 组合框 (settable, string) 当前项目 Value: starter-project
22 按钮 打开项目 Description: 打开项目
23 文本 D:\Ai-project\REA逆向\Agent-Workbench\releases\Agent-Workbench-0.1.0-win-x64\examples\starter-project
24 文本 任务会话
25 按钮 新建任务 Description: 新建任务
26 按钮 新建任务 ↵
27 按钮 创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D
28 按钮 归档会话 创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D Description: 归档会话 创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D
29 按钮 了解这个项目,并提出一项改进
30 按钮 归档会话 了解这个项目,并提出一项改进 Description: 归档会话 了解这个项目,并提出一项改进
31 文本 项目文件
32 文本 5
33 按钮 刷新文件 Description: 刷新文件
34 编辑 (settable, string) 筛选文件 Description: 筛选文件
35 文本 ⌃ K
90 按钮 src Description: src
91 按钮 catalog.js Description: src/catalog.js
92 按钮 tests Description: tests
93 按钮 catalog.test.mjs Description: tests/catalog.test.mjs
94 按钮 package.json Description: package.json
95 按钮 pelican-cycle.html Description: pelican-cycle.html
96 按钮 README.md Description: README.md
36 按钮 机制研究与设计
37 按钮 项目记忆
38 按钮 运行诊断
39 按钮 选择与管理模型
40 文本 本地服务已连接
41 按钮 版本与更新
42 主要
43 章节标题
44 文本 starter-project
45 文本 任务
46 标题 创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D
47 按钮 手机连接
48 按钮 桌面宿主
49 按钮 打开模型中心
50 按钮 切换当前模型 Description: 选择模型、模型池或添加连接
51 文本 本地项目
97 文本 已完成
53 按钮 运行诊断 Description: 运行诊断
54 按钮 项目技能 Description: 项目技能
55 按钮 项目MCP工具 Description: 项目MCP工具
56 按钮 查看任务记录 Description: 查看任务记录
57 按钮 隔离协作 Description: 隔离协作
58 按钮 工程评测 Description: 工程评测
59 按钮 本地代理引擎 Description: 本地代理引擎
60 按钮 REA 工件研究 Description: REA 工件研究
98 按钮 导出会话 Description: 导出会话
61 区域 任务对话
99 文本 你
100 文本 你的任务
101 time
102 文本 20:08
103 文本 创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D动画,无需测试。
104 按钮 模型响应已收到 模型服务 5.6s
105 文本 AW
106 文本 编程助手
107 time
108 文本 20:09
109 文本 我会制作一个可直接在浏览器打开的独立 HTML,用 SVG 绘制鹈鹕、自行车和沿途风景,加入踩踏、车轮转动与背景移动动画。按你的要求,不运行测试。
110 按钮 识别工程与Git基线 0.1s
111 按钮 模型连接失败 模型服务 · 第 2 次 60.0s
112 组 模型输出片段
113 文本 输出未完成
114 文本 Chat Completions
115 文本 ·
116 文本 0
117 文本 字符
118 文本 · 1 个工具参数片段
119 文本 片段仅供查看。完整响应核对前不执行工具;中断后保留原记录。
120 文本 ·
121 文本 工作台
122 time
123 文本 20:11
124 警报
125 文本 模型连接
126 文本 模型服务响应超时
127 文本 模型服务响应超时。项目文件和已完成的工具记录已保留。 已接收的流片段保留,请求不会自动重发。
128 文本 可以重试上次模型请求;若持续失败,在模型连接中检查服务地址和连接状态。
129 文本 发生在请求模型响应时
130 文本 1
131 文本 次请求
132 详细信息
133 开合三角标记 查看错误详情
134 文本 查看错误详情
135 按钮 查看运行诊断
136 文本 ·
137 文本 工作台
138 time
139 文本 20:47
140 文本 从已保存的执行记录继续,已完成的工具结果保留。
141 按钮 模型连接失败 模型服务 · 第 2 次 60.0s
142 组 模型输出片段
143 文本 输出未完成
144 文本 Chat Completions
145 文本 ·
146 文本 0
147 文本 字符
148 文本 · 1 个工具参数片段
149 文本 片段仅供查看。完整响应核对前不执行工具;中断后保留原记录。
150 文本 ·
151 文本 工作台
152 time
153 文本 20:49
154 警报
155 文本 模型连接
156 文本 模型服务响应超时
157 文本 模型服务响应超时。项目文件和已完成的工具记录已保留。 已接收的流片段保留,请求不会自动重发。
158 文本 可以重试上次模型请求;若持续失败,在模型连接中检查服务地址和连接状态。
159 文本 发生在请求模型响应时
160 文本 1
161 文本 次请求
162 详细信息
163 开合三角标记 查看错误详情
164 文本 查看错误详情
165 按钮 查看运行诊断
166 文本 ·
167 文本 工作台
168 time
169 文本 21:03
170 文本 从已保存的执行记录继续,已完成的工具结果保留。
171 按钮 模型连接失败 模型服务 · 第 2 次 60.0s
172 组 模型输出片段
173 文本 输出未完成
174 文本 Chat Completions
175 文本 ·
176 文本 0
177 文本 字符
178 文本 · 1 个工具参数片段
179 文本 片段仅供查看。完整响应核对前不执行工具;中断后保留原记录。
180 文本 ·
181 文本 工作台
182 time
183 文本 21:05
184 警报
185 文本 模型连接
186 文本 模型服务响应超时
187 文本 模型服务响应超时。项目文件和已完成的工具记录已保留。 已接收的流片段保留,请求不会自动重发。
188 文本 可以重试上次模型请求;若持续失败,在模型连接中检查服务地址和连接状态。
189 文本 发生在请求模型响应时
190 文本 1
191 文本 次请求
192 详细信息
193 开合三角标记 查看错误详情
194 文本 查看错误详情
195 按钮 查看运行诊断
196 文本 ·
197 文本 工作台
198 time
199 文本 21:26
200 文本 从已保存的执行记录继续,已完成的工具结果保留。
201 按钮 模型响应已收到 模型服务 121.1s
202 按钮 准备文件改动 pelican-cycle.html 0.0s
203 文本 整理请求上下文
204 文本 完整请求输入估算 13568 → 8741 tokens;省略 1 个完整消息组,原历史和参数已保存为SHA256工件。估算与服务实际用量分别记录。
205 按钮 模型响应已收到 模型服务 8.6s
206 文本 AW
207 文本 编程助手
208 time
209 文本 21:28
210 文本 已准备好
211 文本 pelican-cycle.html
212 文本 的文件提案,内容为 SVG 绘制的鹈鹕骑自行车 2D 动画。
213 文本 请在改动面板查看差异并点击“应用”;应用后即可用浏览器打开该 HTML。当前提案尚未写入项目。
214 文本 按你的要求,未运行测试。
215 文本 ·
216 文本 工作台
217 time
218 文本 21:28
219 文本 文件已审查并应用:pelican-cycle.html。按照原要求未运行测试,可在文件列表打开结果。
220 文本 已应用 pelican-cycle.html
221 文本 写入前基线检查通过,已创建新文件。
69 章节页脚
70 按钮 添加截图
71 文本 每任务1张 · PNG/JPEG · 原图≤2MB
72 编辑 (settable, string) 任务内容 Description: 描述项目任务,或询问工作台的执行状态与报错…
73 文本 starter-project
74 文本 高性能
75 按钮 (disabled) 发送任务
76 文本 Ctrl + Enter 发送 · 会话自动保存
77 文本 上下文估算 8,741 tokens
78 补充
79 缩略图 调整检视区宽度 Value: 50
80 按钮 文件
81 按钮 改动
82 按钮 轨迹
83 按钮 任务
84 按钮 索引
85 按钮 浏览
86 按钮 放大检视区 Description: 放大检视区
87 标题 查看项目文件
88 文本 从左侧打开文件,或让助手先读取项目。
89 文本 读取范围限定在当前项目
223 对话框 模型中心 Secondary Actions: Raise
224 标题 模型中心
225 文本 连接、模型池与执行健康统一管理。
226 按钮 关闭窗口 Description: 关闭窗口
227 区域 模型控制台
228 文本 MODEL CONTROL
229 标题 选择模型,继续开发。
230 文本 连接、模型池和实际请求使用同一个入口。
231 文本 新任务默认
232 文本 高性能
233 文本 0
234 文本 等待
235 导航 模型中心分类
236 按钮 模型连接
237 按钮 连接诊断
238 按钮 模型池
239 按钮 运行预算
240 按钮 等待时限
241 按钮 请求账本
242 按钮 配置导入
243 按钮 凭据保存
244 按钮 运行巡检
245 按钮 刷新模型中心
247 区域 默认模型选择
248 标题 新任务默认模型
249 文本 选择已保存的连接,或使用模型池调度。切换本身不请求模型。
250 按钮 添加模型
251 文本 当前选择
252 组合框 (settable, string) 新任务默认模型 Value: 高性能 · 2 个连接
253 文本 模型池按成员配置调度;每次实际连接和返回型号保存在请求账本。
254 文本 当前会话最近运行
255 文本 jojocode · gpt-6.1-sol;apikey · claude-opus-5-5
256 文本 这是该次运行封存的请求配置;实际返回型号请看请求账本。更改默认选择不会改写旧记录。
257 标题 模型连接
258 文本 保存多条连接后可在上方切换;添加连接不会替换当前模型。凭据可选择本机加密保存。
259 按钮 添加连接
260 文章
261 标题 原模型连接
262 文本 兼容服务
263 文本 ·
264 文本 Chat Completions
265 文本 已配置
266 列表
267 字词 模型
268 文本 模型
269 文本 gpt-6.1-sol
270 字词 请求地址
271 文本 请求地址
272 文本 https://jojocode.com/v1/chat/completions
273 字词 凭据
274 文本 凭据
275 文本 已设置,重启需重新填写
276 字词 请求 / 容量
277 文本 请求 / 容量
278 文本 0
279 文本 在途 · 并发
280 文本 1
281 文本 · 上下文
282 文本 16,000
283 文本 文本
284 文本 :
285 文本 声明支持
286 文本 工具
287 文本 :
288 文本 声明支持
289 文本 流式
290 文本 :
291 文本 待验证
292 文本 视觉
293 文本 :
294 文本 待验证
295 文本 语音
296 文本 :
297 文本 待验证
298 章节页脚
299 按钮 用于新任务
300 按钮 编辑
301 按钮 连接诊断
302 文本 检测范围
303 组合框 (settable, string) 原模型连接检测范围 Value: 文本响应
304 按钮 原生参数
305 按钮 验证响应(调用模型)
306 文章
307 标题 apikey
308 文本 GPT 兼容代理
309 文本 ·
310 文本 Responses
311 文本 · 流式开启
312 文本 已停用
313 列表
314 字词 模型
315 文本 模型
316 文本 gpt-6.1-sol
317 字词 请求地址
318 文本 请求地址
319 文本 https://api.apikey.fan/v1/responses
320 字词 凭据
321 文本 凭据
322 文本 已设置,重启需重新填写
323 字词 请求 / 容量
324 文本 请求 / 容量
325 文本 0
326 文本 在途 · 并发
327 文本 8
328 文本 · 上下文
329 文本 16,000
330 文本 文本
331 文本 :
332 文本 声明支持
333 文本 工具
334 文本 :
335 文本 声明支持
336 文本 流式
337 文本 :
338 文本 声明支持
339 文本 视觉
340 文本 :
341 文本 待验证
342 文本 语音
343 文本 :
344 文本 待验证
345 章节页脚
346 按钮 (disabled) 用于新任务
347 按钮 编辑
348 按钮 连接诊断
349 按钮 启用apikey
350 文本 检测范围
351 组合框 (settable, string) apikey检测范围 Value: 文本响应
352 按钮 原生参数
353 按钮 (disabled) 验证响应(调用模型)
354 文章
355 标题 apikey
356 文本 OpenAI
357 文本 ·
358 文本 Chat Completions
359 文本 · 流式开启
360 文本 已配置
361 列表
362 字词 模型
363 文本 模型
364 文本 gpt-6.1-sol
365 字词 请求地址
366 文本 请求地址
367 文本 https://api.apikey.fan/v1/chat/completions
368 字词 凭据
369 文本 凭据
370 文本 已设置,重启需重新填写
371 字词 请求 / 容量
372 文本 请求 / 容量
373 文本 0
374 文本 在途 · 并发
375 文本 1
376 文本 · 上下文
377 文本 16,000
378 文本 文本
379 文本 :
380 文本 声明支持
381 文本 工具
382 文本 :
383 文本 声明支持
384 文本 流式
385 文本 :
386 文本 声明支持
387 文本 视觉
388 文本 :
389 文本 待验证
390 文本 语音
391 文本 :
392 文本 待验证
393 章节页脚
394 按钮 用于新任务
395 按钮 编辑
396 按钮 连接诊断
397 按钮 停用apikey
398 文本 检测范围
399 组合框 (settable, string) apikey检测范围 Value: 文本响应
400 按钮 原生参数
401 按钮 验证响应(调用模型)
402 文章
403 标题 jojocode
404 文本 兼容服务
405 文本 ·
406 文本 Chat Completions
407 文本 已配置
408 列表
409 字词 模型
410 文本 模型
411 文本 gpt-6.1-sol
412 字词 请求地址
413 文本 请求地址
414 文本 https://jojocode.com/v1/chat/completions
415 字词 凭据
416 文本 凭据
417 文本 已设置,重启需重新填写
418 字词 请求 / 容量
419 文本 请求 / 容量
420 文本 0
421 文本 在途 · 并发
422 文本 8
423 文本 · 上下文
424 文本 16,000
425 文本 文本
426 文本 :
427 文本 声明支持
428 文本 工具
429 文本 :
430 文本 声明支持
431 文本 流式
432 文本 :
433 文本 待验证
434 文本 视觉
435 文本 :
436 文本 待验证
437 文本 语音
438 文本 :
439 文本 待验证
440 章节页脚
441 按钮 用于新任务
442 按钮 编辑
443 按钮 连接诊断
444 按钮 停用jojocode
445 文本 检测范围
446 组合框 (settable, string) jojocode检测范围 Value: 文本响应
447 按钮 原生参数
448 按钮 验证响应(调用模型)
12 区域
Document text: ```
AW
Agent Workbench
本地编程工作台 · 正式版
收起项目栏
当前项目
当前项目
打开项目
D:\Ai-project\REA逆向\Agent-Workbench\releases\Agent-Workbench-0.1.0-win-x64\examples\starter-projectD:\Ai-project\REA逆向\Agent-Wor…
任务会话
新建任务
新建任务
↵
创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D创建一个HTML,内容是SVG…
归档会话 创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D
了解这个项目,并提出一项改进了解这个项目,并提出一项…
归档会话 了解这个项目,并提出一项改进
项目文件 5
刷新文件
筛选文件
⌃ K
src
catalog.js
tests
catalog.test.mjs
package.json
pelican-cycle.html
README.md
机制研究与设计
项目记忆
运行诊断
模型与连接
高性能
本地服务已连接
v1.24.7
starter-project任务
创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D创建一个HTML,内容是SVG绘制一…
手机连接
桌面宿主
模型中心
高性能
本地项目
已完成
运行诊断
项目技能
项目MCP工具
查看任务记录
隔离协作
工程评测
本地代理引擎
REA 工件研究
导出会话
你
你的任务
20:08
创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D动画,无需测试。
模型响应已收到
模型服务
5.6s
AW
编程助手
20:09
我会制作一个可直接在浏览器打开的独立 HTML,用 SVG 绘制鹈鹕、自行车和沿途风景,加入踩踏、车轮转动与背景移动动画。按你的要求,不运行测试。
识别工程与Git基线
0.1s
模型连接失败
模型服务 · 第 2 次
60.0s
输出未完成
Chat Completions · 0 字符 · 1 个工具参数片段
片段仅供查看。完整响应核对前不执行工具;中断后保留原记录。
·
工作台
20:11
模型连接
模型服务响应超时
模型服务响应超时。项目文件和已完成的工具记录已保留。 已接收的流片段保留,请求不会自动重发。
可以重试上次模型请求;若持续失败,在模型连接中检查服务地址和连接状态。
发生在请求模型响应时
1 次请求
查看错误详情
查看运行诊断
·
工作台
20:47
从已保存的执行记录继续,已完成的工具结果保留。
模型连接失败
模型服务 · 第 2 次
60.0s
输出未完成
Chat Completions · 0 字符 · 1 个工具参数片段
片段仅供查看。完整响应核对前不执行工具;中断后保留原记录。
·
工作台
20:49
模型连接
模型服务响应超时
模型服务响应超时。项目文件和已完成的工具记录已保留。 已接收的流片段保留,请求不会自动重发。
可以重试上次模型请求;若持续失败,在模型连接中检查服务地址和连接状态。
发生在请求模型响应时
1 次请求
查看错误详情
查看运行诊断
·
工作台
21:03
从已保存的执行记录继续,已完成的工具结果保留。
模型连接失败
模型服务 · 第 2 次
60.0s
输出未完成
Chat Completions · 0 字符 · 1 个工具参数片段
片段仅供查看。完整响应核对前不执行工具;中断后保留原记录。
·
工作台
21:05
模型连接
模型服务响应超时
模型服务响应超时。项目文件和已完成的工具记录已保留。 已接收的流片段保留,请求不会自动重发。
可以重试上次模型请求;若持续失败,在模型连接中检查服务地址和连接状态。
发生在请求模型响应时
1 次请求
查看错误详情
查看运行诊断
·
工作台
21:26
从已保存的执行记录继续,已完成的工具结果保留。
模型响应已收到
模型服务
121.1s
准备文件改动
pelican-cycle.html
0.0s
整理请求上下文
完整请求输入估算 13568 → 8741 tokens;省略 1 个完整消息组,原历史和参数已保存为SHA256工件。估算与服务实际用量分别记录。
模型响应已收到
模型服务
8.6s
AW
编程助手
21:28
已准备好 pelican-cycle.html 的文件提案,内容为 SVG 绘制的鹈鹕骑自行车 2D 动画。
请在改动面板查看差异并点击“应用”;应用后即可用浏览器打开该 HTML。当前提案尚未写入项目。
按你的要求,未运行测试。
·
工作台
21:28
文件已审查并应用:pelican-cycle.html。按照原要求未运行测试,可在文件列表打开结果。
已应用 pelican-cycle.html
写入前基线检查通过,已创建新文件。
添加截图
每任务1张 · PNG/JPEG · 原图≤2MB
任务内容
starter-project
高性能
发送任务
Ctrl + Enter 发送 · 会话自动保存
上下文估算 8,741 tokens
调整检视区宽度
文件
改动
轨迹
任务
索引
浏览
放大检视区

查看项目文件
从左侧打开文件,或让助手先读取项目。
读取范围限定在当前项目
模型中心
连接、模型池与执行健康统一管理。
关闭窗口
MODEL CONTROL
选择模型,继续开发。
连接、模型池和实际请求使用同一个入口。
新任务默认
高性能
0 等待
模型连接
连接诊断
模型池
运行预算
等待时限
请求账本
配置导入
凭据保存
运行巡检
刷新模型中心
新任务默认模型
选择已保存的连接,或使用模型池调度。切换本身不请求模型。
添加模型
当前选择
新任务默认模型
模型池按成员配置调度;每次实际连接和返回型号保存在请求账本。
当前会话最近运行
jojocode · gpt-6.1-sol;apikey · claude-opus-5-5
这是该次运行封存的请求配置;实际返回型号请看请求账本。更改默认选择不会改写旧记录。
模型连接
保存多条连接后可在上方切换;添加连接不会替换当前模型。凭据可选择本机加密保存。
添加连接
原模型连接
兼容服务 · Chat Completions
已配置
模型
gpt-6.1-sol
请求地址
https://jojocode.com/v1/chat/completions
凭据
已设置,重启需重新填写
请求 / 容量
0 在途 · 并发 1 · 上下文 16,000
文本:声明支持
工具:声明支持
流式:待验证
视觉:待验证
语音:待验证
用于新任务
编辑
连接诊断
检测范围
原模型连接检测范围
原生参数
验证响应(调用模型)
apikey
GPT 兼容代理 · Responses · 流式开启
已停用
模型
gpt-6.1-sol
请求地址
https://api.apikey.fan/v1/responses
凭据
已设置,重启需重新填写
请求 / 容量
0 在途 · 并发 8 · 上下文 16,000
文本:声明支持
工具:声明支持
流式:声明支持
视觉:待验证
语音:待验证
用于新任务
编辑
连接诊断
启用
检测范围
apikey检测范围
原生参数
验证响应(调用模型)
apikey
OpenAI · Chat Completions · 流式开启
已配置
模型
gpt-6.1-sol
请求地址
https://api.apikey.fan/v1/chat/completions
凭据
已设置,重启需重新填写
请求 / 容量
0 在途 · 并发 1 · 上下文 16,000
文本:声明支持
工具:声明支持
流式:声明支持
视觉:待验证
语音:待验证
用于新任务
编辑
连接诊断
停用
检测范围
apikey检测范围
原生参数
验证响应(调用模型)
jojocode
兼容服务 · Chat Completions
已配置
模型
gpt-6.1-sol
请求地址
https://jojocode.com/v1/chat/completions
凭据
已设置,重启需重新填写
请求 / 容量
0 在途 · 并发 8 · 上下文 16,000
文本:声明支持
工具:声明支持
流式:待验证
视觉:待验证
语音:待验证
用于新任务
编辑
连接诊断
停用
检测范围
jojocode检测范围
原生参数
验证响应(调用模型)
```
@@ -0,0 +1,18 @@
{
"success": true,
"version": "1.24.7",
"pid": 36552,
"desktopWindows": 1,
"runtimeRoot": "C:\\Users\\wangy\\AppData\\Local\\Programs\\AgentWorkbench\\versions\\Agent-Workbench-1.24.7-win-x64-20261103164618190",
"observed": [
"Actual formal installed launcher opens unique 1.24.7 native window",
"Connected UI and original high-performance pool visible",
"No model dispatch, settings save, task start or original history mutation"
],
"screenshots": [
"after.jpg",
"model-center.jpg"
],
"realModelRequests": 0,
"scope": "Computer Use observation from the actual installed original entry, not Electron fixture. Original R budget behavior is separately verified by write-blocked production web UI. State reconciliation follows after native review."
}
@@ -0,0 +1,84 @@
{
"portableRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.24.7-win-x64-20261103164618190",
"sourceRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.24.7-source-20261103164618190",
"evidenceRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.24.7-evidence-20261103164618190",
"portable": {
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.24.7-win-x64-20261103164618190.zip",
"bytes": 242883511,
"sha256": "9f6ec8ae971d9207b34d5a26c2b8e3c09e63b3138793ef25d898ce6e3641b23a"
},
"source": {
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.24.7-source-20261103164618190.zip",
"bytes": 1020471144,
"sha256": "e27ae51497a2a42632486a4b852e98a32c2d0877e9874f85ad1a01b802f49f1e"
},
"evidence": {
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.24.7-evidence-20261103164618190.zip",
"bytes": 779265448,
"sha256": "064f6bd47d6900f63706300ce1594b859ef0a92eebb41f6d656b386ad2981291"
},
"documentation": {
"sourceCommit": "5dcd6d7c3f4f465071ae1156bb1e0bf5a28f48e2",
"originalFiles": 20728,
"originalBytes": 2443917643,
"shippedFiles": 154,
"shippedBytes": 1170702,
"evidenceManifestSha256": "dba6ac06d50083cfcc9724f5bd56a2befd0f1e69eccd3208c58e63632472f360"
},
"installer": {
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.24.7-win-x64-20261103164618190-Setup.exe",
"bytes": 213284660,
"sha256": "a2ea5ce3e249f5aeecc7bb5a8972746a214a59343d95d5828d6a0707a67dc2c2",
"compiler": {
"version": "6.3.1",
"sha256": "3b22d36fc7e7c421d8caf780008ecdb42b14b3f0371b9bafe3cad64e6648d9cf"
},
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\installer-build-2bda8bad-f0af-4582-8304-200b8e7a1078\\installer.iss",
"version": "1.24.7",
"candidateDirectory": "Agent-Workbench-1.24.7-win-x64-20261103164618190",
"qaFixture": null
},
"desktopRuntime": {
"version": "45.2.0",
"sha256": "010ea51bfb528c8143513439aa93fb40754403daa2023f9decd4517da71706ab",
"bytes": 249810432,
"executable": "desktop/AgentWorkbench.exe",
"scope": "Fixed original Windows Electron runtime, full license texts, isolated application loader and native bootstrap; original executable hash unchanged."
},
"nativeLauncher": {
"bytes": 12288,
"sha256": "e3d31cb7ce5009d3ca5a96d04dc16b1add0bd1f5cde7e1304461de6713d37463",
"inputs": {
"compiler": "46809206887326d2d24db1eff1f3064de972c3451abe766b49111450a5e08e00",
"source": "1c8e7d7ee5c383897bec48e9aadb8051b47c09686f74bbf34e0b6b695cd01ae5",
"icon": "f1879074e45f01c480ad9bf50b818637cc60ad56526056c886f808a4812d7b0c"
},
"compiler": "Windows .NET Framework C# compiler; source-bound cached original output, no deterministic-build claim."
},
"nodeVersion": "v24.22.1",
"reaRuntime": {
"version": "6.8.0",
"protocolVersion": "2025-03-26",
"packages": 149,
"files": 8413,
"bytes": 142395402,
"lockSha256": "687fb3fa3089a12f239dd5e87d9fa07df55a87df4935931d495a92600a1b0f6e"
},
"portableFiles": 10949,
"licenseTexts": 135,
"browserRuntime": {
"playwrightVersion": "1.64.1",
"packages": [
{
"name": "playwright",
"version": "1.64.1",
"files": 60
},
{
"name": "playwright-core",
"version": "1.64.1",
"files": 115
}
]
}
}

Some files were not shown because too many files have changed in this diff Show More