docs: seal 1.20.3 local deployment and browser roadmap

This commit is contained in:
Agent Workbench Dev
2026-11-01 23:00:50 +08:00
parent 4bf4f7a8b2
commit 9961d88d59
119 changed files with 8886 additions and 13 deletions
+3
View File
@@ -63,3 +63,6 @@ docs/engineering-evals-*.json -text whitespace=blank-at-eol,blank-at-eof,space-b
docs/model-wait-release-evidence/** -text whitespace=blank-at-eol,blank-at-eof,space-before-tab,cr-at-eol
docs/model-wait-release-identity.json -text whitespace=blank-at-eol,blank-at-eof,space-before-tab,cr-at-eol
docs/model-fidelity-release-evidence/** -text whitespace=blank-at-eol,blank-at-eof,space-before-tab,cr-at-eol
docs/model-fidelity-release-identity.json -text whitespace=blank-at-eol,blank-at-eof,space-before-tab,cr-at-eol
+5 -3
View File
@@ -1,12 +1,14 @@
# Agent Workbench项目执行规范
1.20.2已完成模型响应等待修复与保留状态部署:单次响应默认300秒、流式无数据120秒,模型中心可调整。111项专项/关联回归、真实本机JSON/SSE各65秒、五组等待UI/五组协议UI及五组随包等待UI通过;实际旧版到新版升级及正式Edge/Electron两视口回看通过。正式PID36552/Provider、17项目/31全部会话/3207记录/6138事件和1761个工件文件原字节保持,仅公开模型中心增加计算出的等待默认值;磁盘配置不回写。本轮验收发起真实模型请求0。 当前身份以docs/model-wait-release-identity.json为准,上一版记录保持历史含义。
1.20.3已完成模型原生参数与错误保留的本地部署:HTTP/SSE上游原文、code/type/请求ID、原Retry-After、真实用量与来源可核对;现有Chat/Responses原生参数逐字段发送。最终17项专项、相关103项、串行升级2项/安装10项/Python13项分别通过;源码与最终随包各六组/十截图、实际旧版升级和正式Edge/Electron两视口通过。原PID36552/Provider、原连接/31个原会话及原执行记录/导出/工件保持;激活校验期间新增1会话、16记录、57事件、9工件及2次用户模型请求,已严格核对所属和全部原字节,完整当前17项目/32全部会话/3239记录/6229事件和1780个工件文件再次比较一致,9229关闭。本轮验收发起真实模型请求0,用户自行执行的2次另记。广泛源码原轮646/649,三项验收环境失败与独立复验均保留,不宣称最终冻结源码完整整轮通过。 当前身份以docs/model-fidelity-release-identity.json为准,1.21工程环境和1.22内置浏览器仍为计划。
1.20.2已完成模型响应等待修复与保留状态部署:单次响应默认300秒、流式无数据120秒,模型中心可调整。111项专项/关联回归、真实本机JSON/SSE各65秒、五组等待UI/五组协议UI及五组随包等待UI通过;实际旧版到新版升级及正式Edge/Electron两视口回看通过。正式PID36552/Provider、17项目/31全部会话/3207记录/6138事件和1761个工件文件原字节保持,仅公开模型中心增加计算出的等待默认值;磁盘配置不回写。本轮验收发起真实模型请求0。 该历史版身份以docs/model-wait-release-identity.json为准,原记录保持历史含义。
1.20.1已完成真实案例修复及保留状态本地升级:原任务目标/无需测试约定、只读诊断、系统审查来源、完整无参协议兼容、失败请求用量、无进展保护和实际模型展示均已修复,并补齐暂停封存及Windows进程退出竞态。独立源码626/626、桌面/390px核心5组、相关历史UI5类、同载荷安装/重启/卸载7组、正式26页和实际Electron两视口通过。原PID36552/Provider及模型连接保持,17项目/30全部会话/3144记录/5986事件、全部表每行/全导出/模型中心及1824工件记录/1721去重文件原字节保持,9229关闭。本轮真实模型请求0,未续跑原用户任务或改写历史用量。 不可变Agent-Workbench-1.20.1-win-x64-20261101064603710,实现f3dd39e522c45ed5311ce2cf14c00ef8e0510be8;原件见docs/task-control-j-release-evidence/manifest.json和docs/task-control-release-identity.json。
## 范围与目标
当前接续优先级:用户要求“先修复系统”已完成1.20.1案例A–F及1.20.2当前页面超时修复、相关验收和保留状态部署。已有文件复用交付边界另记后续。1.21与安装进度/性能计划保留;本轮不续跑原react-source任务、重跑其检查、主动使用真实模型或自动启动新功能。后续按最新用户要求及docs/NEXT-STEPS.md推进。
当前接续优先级:1.20.3模型原生参数与错误保留已本地上线,原状态/失败保持;内置浏览器已排1.22。历史用户要求“先修复系统”已完成1.20.1案例A–F及1.20.2当前页面超时修复、相关验收和保留状态部署。已有文件复用交付边界另记后续。1.21与安装进度/性能计划保留;本轮不续跑原react-source任务、重跑其检查、主动使用真实模型或自动启动新功能。后续按最新用户要求及docs/NEXT-STEPS.md推进。
本仓库专注AI编程与能力研究工作台。其他项目不纳入本项目计划或代码。用户确定的研究对象为Codex、Claude Code、Cursor、DeepSeek Harness、WorkBuddy及DSH-purge;目标是将有证据的机制独立开发成自有Harness内置能力。
@@ -64,7 +66,7 @@
## 当前本地服务与Windows
- 正式服务入口http://127.0.0.1:4318,当前1.20.2,沿用原PID36552与Provider;当前身份以docs/model-wait-release-identity.json为准。原模型连接、全部表/会话/工件保持;普通源码/规划/Git工作不重启服务、不读取或复制Key。
- 正式服务入口http://127.0.0.1:4318,当前1.20.3,沿用原PID36552与Provider;当前身份以docs/model-fidelity-release-identity.json为准。原模型连接、全部表/会话/工件保持;普通源码/规划/Git工作不重启服务、不读取或复制Key。
- 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。
+2 -2
View File
@@ -1,8 +1,8 @@
# 本地 Agent 工作台
1.20.3新增上游错误详情/原文下载、原生参数配置和按原因处理429,见[验收记录](docs/QA-1.20.3.md)。正式部署以最终发行身份为准。内置浏览器已明确为[1.22阶段](docs/PLAN-1.22.md),当前已有静态工程页面验收。
1.20.3已本地上线:上游错误详情/原文下载、原生参数配置和按原因处理429。见[工作台](http://127.0.0.1:4318/?release=1.20.3)、[验收](docs/QA-1.20.3.md)、[回顾](docs/RETROSPECTIVE-1.20.3.md)和[发行身份](docs/model-fidelity-release-identity.json)。内置浏览器已明确为[1.22阶段](docs/PLAN-1.22.md),当前已有静态工程页面验收。
上一维护版1.20.2修复固定60秒生成截止,新增模型中心「等待时限」;当前正式版本已升级,原连接和历史保持。见[工作台](http://127.0.0.1:4318/?release=1.20.2)、[验收](docs/QA-1.20.2.md)、[回顾](docs/RETROSPECTIVE-1.20.2.md)和[发行身份](docs/model-wait-release-identity.json)。历史失败仍按原记录显示,手动续接的新请求使用新时限,任务总预算仍优先。
上一维护版1.20.2修复固定60秒生成截止,新增模型中心「等待时限」;其部署历史及原连接/记录保持。见[工作台](http://127.0.0.1:4318/?release=1.20.2)、[验收](docs/QA-1.20.2.md)、[回顾](docs/RETROSPECTIVE-1.20.2.md)和[发行身份](docs/model-wait-release-identity.json)。历史失败仍按原记录显示,手动续接的新请求使用新时限,任务总预算仍优先。
1.20.1已修复任务目标丢失、审查后附加测试、完整无参工具报错、失败消费漏记和无进展循环,并修复暂停收尾及Windows退出竞态。独立626项、实际桌面/390px、安装/升级通过,原连接与全部历史保持。见[工作台](http://127.0.0.1:4318/?release=1.20.1)、[验收](docs/QA-1.20.1.md)、[回顾](docs/RETROSPECTIVE-1.20.1.md)和[发行身份](docs/task-control-release-identity.json)。1.21计划保留,真实模型自主工程复验单列。
+4 -1
View File
@@ -1,6 +1,6 @@
# 模型能力与错误信息保留方案
用户提出工作台应保留大模型能力,错误也应传到用户面前,并询问Codex是否为原样透传。本记录固定接入原则与已确认缺口。用户随后授权实现上线,1.20.3候选已完成HTTP/SSE错误、重试语义、错误界面与现有Chat/Responses原生参数接入;最终验证/部署见[QA-1.20.3](QA-1.20.3.md),不将候选阶段写为正式上线。更多原生参数与新模态继续逐项开发,原会话不回写。
用户提出工作台应保留大模型能力,错误也应传到用户面前,并询问Codex是否为原样透传。本记录固定接入原则与已确认缺口。用户随后授权实现上线,1.20.3已完成HTTP/SSE错误、重试语义、错误界面与现有Chat/Responses原生参数接入及本地部署;实际范围、原失败与验证见[QA-1.20.3](QA-1.20.3.md)。更多原生参数与新模态继续逐项开发,原会话不回写。
## 接入原则
@@ -37,3 +37,6 @@
[OpenAI官方架构文档](https://developers.openai.com/api/docs/guides/agents-api/architecture#the-pieces)将Harness定义为运行模型与工具循环、维护会话的执行系统,并区分实际执行环境。该页描述托管Agents API,不能据此证明当前桌面应用的每个内部网络字段均被原样展示。当前所见429和request ID也不能独自证明全部错误元数据完整透传。
[OpenAI官方错误文档](https://developers.openai.com/api/docs/guides/error-codes#api-errors)区分429请求速率、额度、组织/项目消费及用量限制;需要具体error.code才能对应处理。用户本次429明确出现在Codex对话,属于与工作台本机超时不同的请求链。
1.20.3已完成本地部署及保留状态验收,实际范围/原失败/后续边界以[验收](QA-1.20.3.md)、[回顾](RETROSPECTIVE-1.20.3.md)与[发行身份](model-fidelity-release-identity.json)为准;更多模态与供应商字段不因此视为已接入。
+2 -2
View File
@@ -1,8 +1,8 @@
# 后续执行计划
最新执行顺序:1.20.3实现并本地上线模型错误保留与现有协议原生参数 → [1.21工程依赖与可复现环境](PLAN-1.21.md) → [1.22工作台内置浏览器](PLAN-1.22.md)。1.10已有静态构建的真实Edge页面验收,1.22才是项目开发预览、桌面内嵌标签页与可观察/停止的通用网站操作;Codex右侧浏览器不算本产品内置能力。
最新执行顺序:1.20.3模型错误保留与现有协议原生参数已本地上线 → [1.21工程依赖与可复现环境](PLAN-1.21.md) → [1.22工作台内置浏览器](PLAN-1.22.md)。1.10已有静态构建的真实Edge页面验收,1.22才是项目开发预览、桌面内嵌标签页与可观察/停止的通用网站操作;Codex右侧浏览器不算本产品内置能力。
新增接入要求按[模型能力与错误信息保留方案](MODEL-FIDELITY-PLAN.md)执行:1.20.3候选已实现有界HTTP/SSE原错误、上游request ID、原code/type与Retry-After及现有Chat/Responses原生参数;实际验证/部署以[QA-1.20.3](QA-1.20.3.md)为准。更多供应商参数与模态仍逐项验证,不能称为全能力透传。原失败不回写,本机ID与上游ID分开,不把中文解释当原文。
新增接入要求按[模型能力与错误信息保留方案](MODEL-FIDELITY-PLAN.md)执行:1.20.3已本地交付有界HTTP/SSE原错误、上游request ID、原code/type与Retry-After及现有Chat/Responses原生参数;实际验证/部署以[QA-1.20.3](QA-1.20.3.md)为准。更多供应商参数与模态仍逐项验证,不能称为全能力透传。原失败不回写,本机ID与上游ID分开,不把中文解释当原文。
本轮正式暂存及独立核对实际耗时860秒,最终安装指针激活及原状态核对647秒,发行19560文件。原暂存后因新增模型尝试被拒绝的结果保持失败,耗时不能当成已通过的性能指标。后续维护优先研究运行载荷与历史验收材料分离、同一发行的必要核对复用,以及当前校验阶段/耗时/失败收据的可见反馈;保留路径、字节变化和回滚门槛,另行实现验收。
+14 -1
View File
@@ -31,4 +31,17 @@ TypeScript、Vite、手机静态壳及服务端构建通过。Edge两视口六
首个候选20261101140547938已通过362份Git输入、22份编译产物和完整发行身份核对,实际随包六组/十截图通过;独立1.20.2升级时被升级脚本的旧支持版本清单拒绝,尚未打开调试器或改变正式服务。已补齐1.20.2及当前1.20.3来源兼容,原两项真实进程升级测试通过,服务端/安装器重新构建。首候选及原失败保留,新不可变候选另行生成与验收,不能沿用首候选的身份。
待完成候选与保留状态升级,当前此处不登记上线成功。
1.20.3已完成模型原生参数与错误保留的本地部署:HTTP/SSE上游原文、code/type/请求ID、原Retry-After、真实用量与来源可核对;现有Chat/Responses原生参数逐字段发送。最终17项专项、相关103项、串行升级2项/安装10项/Python13项分别通过;源码与最终随包各六组/十截图、实际旧版升级和正式Edge/Electron两视口通过。原PID36552/Provider、原连接/31个原会话及原执行记录/导出/工件保持;激活校验期间新增1会话、16记录、57事件、9工件及2次用户模型请求,已严格核对所属和全部原字节,完整当前17项目/32全部会话/3239记录/6229事件和1780个工件文件再次比较一致,9229关闭。本轮验收发起真实模型请求0,用户自行执行的2次另记。广泛源码原轮646/649,三项验收环境失败与独立复验均保留,不宣称最终冻结源码完整整轮通过。
最终实现来源为`4bf4f7a8b271264420b7450c08a305529570d0dd`,不可变候选为`Agent-Workbench-1.20.3-win-x64-20261101141644779`。362份Git输入/22份编译产物、完整portable清单和source/zip/Setup身份已核对。实际随包六组/十截图通过,独立已安装1.20.2宿主升级到最终包后原PID/所属凭据/池及每表/导出/工件保持。正式Setup暂存、一次保留Provider切换通过,产品指针已激活;外层原基线比较因新增用户活动被拒绝,保留原拒绝后只读核对激活收据和新增所属,最新稳定基线整表/全导出/工件/模型中心精确比较通过;四阶段分别875/9/730/11秒,是阶段墙钟耗时,不是单纯切换或模型推理耗时。
正式Edge与实际安装的可见Electron各桌面/390px,共4个视图/8张截图。指定原会话保持部署基线中的完成状态,原生参数入口按实际连接显示,保存按钮可达;没有提交配置、重试原任务或调用模型,界面无页面错误/被阻写请求。完整历史错误不补写新字段。
广泛源码轮649项646通过、3失败:两项Python因未设独立QA解释器缺pytest,另一项旧宿主升级因并行测试占用9229而被正确拒绝。随后既有独立Python/pytest8.4.2下13/13和升级串行2/2通过;安装10/10及最终专项17/17另列。该广泛轮开始于最后HTTP用量及升级兼容修复前,各次来源和原失败不合并伪装为最终冻结整轮。
原件与SHA见[证据清单](model-fidelity-release-evidence/manifest.json)、[发行身份](model-fidelity-release-identity.json)及[回顾](RETROSPECTIVE-1.20.3.md)。SQLite备份、完整正式快照和用户原导出仅保留在忽略的本机QA目录。现有供应商实际接受参数、真实模型推理及新一轮实机手机验收未执行。
激活期间用户自行创建会话90d02c6d并完成2次模型请求。原31会话、原执行记录、原导出和原工件逐项保持;新增16记录/57事件/9工件均归属于同一新Task/Run,原68个请求保持且只追加2个,唯一使用连接的健康时间随原响应前进。原激活外层拒绝仍保留,未重新安装/激活;以产品原激活收据、严格追加来源证明、最新备份与稳定基线继续只读验收。验收器没有主动调用真实模型,不把用户运行算为本轮合成测试或全面真实推理验收。
正式UI第一轮错误沿用原会话failed假设,但该会话在部署基线已completed;第二轮窄屏命中隐藏侧栏。只修正验收状态和标题定位,原失败与桌面截图保持;第三轮4视图/8截图及全当前状态精确比较通过。
+26
View File
@@ -0,0 +1,26 @@
# 1.20.3 回顾:原生参数与模型错误
1.20.3已完成模型原生参数与错误保留的本地部署:HTTP/SSE上游原文、code/type/请求ID、原Retry-After、真实用量与来源可核对;现有Chat/Responses原生参数逐字段发送。最终17项专项、相关103项、串行升级2项/安装10项/Python13项分别通过;源码与最终随包各六组/十截图、实际旧版升级和正式Edge/Electron两视口通过。原PID36552/Provider、原连接/31个原会话及原执行记录/导出/工件保持;激活校验期间新增1会话、16记录、57事件、9工件及2次用户模型请求,已严格核对所属和全部原字节,完整当前17项目/32全部会话/3239记录/6229事件和1780个工件文件再次比较一致,9229关闭。本轮验收发起真实模型请求0,用户自行执行的2次另记。广泛源码原轮646/649,三项验收环境失败与独立复验均保留,不宣称最终冻结源码完整整轮通过。
## 实际改善
此前HTTP错误只留简短文字,原code/type、上游身份和等待要求不足;SSE失败也容易与本地核验混在一起。现在Task所属错误可核对有界脱敏原文及SHA,并列中文解释、实际连接/物理尝试/上游身份;失败已报告用量继续进入原Run账本。完整性或用量未知时保留未知,不执行片段工具、不重写旧失败。
现有Chat/Responses分别保存明确支持的原生参数,按成员构造实际请求,未知字段在保存/发送前拒绝;原路由和凭据保持。此版没有把任意参数、供应商服务端工具或音视频宣称为已透传。
## 本轮发现与处理
- 首个发行包和随包界面通过,真实旧1.20.2升级仍因版本清单遗漏被拒绝。补齐1.20.2/1.20.3、串行验证重复升级和旧宿主升级,再生成不可变新包;旧包和失败收据保持。这证明来源核对、界面通过及实际升级需要分别验收。
- 广泛源码轮646/649,其中两项验收解释器缺pytest,另一项两个独立升级验证共享9229。原结果保持;既有独立pytest8.4.2环境13项及串行升级2项另轮通过。后续固定运行环境并串行使用固定资源;不将复验拼接成原轮全通过。
- 初次构建缺共享MJS类型声明,首轮新测试夹具混用HTTP/Responses状态,均保留原失败再修复;截图回看发现SSE来源措辞和实际连接上下文被undefined覆盖,已补齐并验证。
## 正式部署与限制
正式四阶段分别875/9/730/11秒,原PID36552/Provider、31个原会话和原工件保持,当前17项目/32全部会话/3239记录/6229事件和1780个工件文件按包含新增活动的稳定基线再次精确比较。实际Edge/Electron两视口只读回看通过,无配置保存、原任务重放或真实模型调用。完整SQLite/用户导出留本机QA,不入Git;验收工件按原字节与Git blob核对。阶段墙钟仍长,既有最小载荷/进度反馈维护项保留,未在此版声称解决。
## 下一步
先按[1.21](PLAN-1.21.md)完成工程依赖诊断、固定准备计划、审批/停止与可复现环境;随后[1.22](PLAN-1.22.md)交付开发预览、桌面标签页、代理页面操作及会话权限。1.10的静态工程Edge验收继续复用,Codex右侧浏览器不计入本产品。已有文件证据复用、更多引擎与语音分别保留,不在当前补丁随意扩展。
激活后的外层比较被新增真实用户会话拒绝;产品指针已经激活,不重复副作用。以原产品收据证明实际完成,以每行/每工件/每导出原子集、所有新增事件/记录/工件/尝试的同一会话归属以及唯一健康时间变化证明合法追加,再作最新稳定基线完整比较。原拒绝不能抹掉,也不因此退回或删除用户新增结果。正式UI脚本两次旧假设/隐藏侧栏定位失败亦保留;修正依据后整轮回看通过。
Binary file not shown.

After

Width:  |  Height:  |  Size: 41 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 89 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 89 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 53 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 46 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 144 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 49 KiB

@@ -0,0 +1,57 @@
{
"success": true,
"appRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-win-x64-20261101141644779",
"checks": [
"native parameters persist after reload and the narrow form remains usable without calling a model",
"unsupported field is rejected visibly before any dispatch",
"actual HTTP error shows raw fields and original wait, keeps secrets hidden and downloads SHA-verified original at both widths",
"physical request ledger preserves the same error origin and offers the sealed actual request",
"Responses failed event retains original identity and reported use at both widths; partial tool is never dispatched or retried",
"explicit synthetic probe also shows the API diagnostic and labels its preview-only retention without inventing a Task artifact"
],
"screenshots": [
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\native-desktop.png",
"sha256": "7eb38b0ccfb87074d9d6011d1b869142b9d0e54ff81ecc951b38845219e7a80f"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\native-390.png",
"sha256": "bf0c6db94447284972e44f196b5136a33a41e71510deec9d5d09202e9290537e"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\http-error-desktop.png",
"sha256": "a26e4de08071acc816cf36d6364aa11cda1c6b46db3945749206c56b4f775ff5"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\http-raw-desktop.png",
"sha256": "772e60beb50e10dc591ea8c726d1554be0af23778e686d3646e5083eb7660570"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\http-error-390.png",
"sha256": "62f389756b0dc1aa9195b62b4b38286ea1e5532f88650e042f282bab68aedea8"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\ledger-390.png",
"sha256": "0db258b881d9a7604b82c600fe264e176858a04e6d16866c63a5fc10587cfb5f"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\sse-error-desktop.png",
"sha256": "9920c5e6318a14031d3f4b92204e3c1ed576dc784af40a76521ef9c090cce0b8"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\sse-error-390.png",
"sha256": "2a56c56c95fcb934c61c4b6cbc3d16b7092bbfb4b48fb49e2a032f378fce9c66"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\sse-raw-390.png",
"sha256": "8bdd5983369d1e2da56f66b81624c6d93025ab4649228b752e82cf725e81dc8c"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\probe-error-390.png",
"sha256": "d90c61454d8665df5e03ea837e2b5143e086c32d1c4ddf57aea38e67f476b263"
}
],
"errors": [],
"modelRequests": 3,
"realModelRequests": 0
}
Binary file not shown.

After

Width:  |  Height:  |  Size: 48 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 93 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 48 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 41 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 89 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 89 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 52 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 46 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 144 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 49 KiB

@@ -0,0 +1,57 @@
{
"success": true,
"appRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench",
"checks": [
"native parameters persist after reload and the narrow form remains usable without calling a model",
"unsupported field is rejected visibly before any dispatch",
"actual HTTP error shows raw fields and original wait, keeps secrets hidden and downloads SHA-verified original at both widths",
"physical request ledger preserves the same error origin and offers the sealed actual request",
"Responses failed event retains original identity and reported use at both widths; partial tool is never dispatched or retried",
"explicit synthetic probe also shows the API diagnostic and labels its preview-only retention without inventing a Task artifact"
],
"screenshots": [
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\native-desktop.png",
"sha256": "7eb38b0ccfb87074d9d6011d1b869142b9d0e54ff81ecc951b38845219e7a80f"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\native-390.png",
"sha256": "bf0c6db94447284972e44f196b5136a33a41e71510deec9d5d09202e9290537e"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\http-error-desktop.png",
"sha256": "67b3d629501ed4cb07917242374f776608938d03cc6d82826d4405ecad967bd2"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\http-raw-desktop.png",
"sha256": "902d3900649c823d8dabe1c2860e82a37efafd455a80a5d15766d905c861ccc7"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\http-error-390.png",
"sha256": "6272836b7ec03d0d5feef464a8023c6c62d0956454e45fae8f41798b57820d46"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\ledger-390.png",
"sha256": "13de4ab0fde40229b41b1e63cbb290adef467ea85310e01cea884f51c36bd4a9"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\sse-error-desktop.png",
"sha256": "db58adc605aaf1126cb1596e2cd28e8891d02c66612dc006c295a23556d646ce"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\sse-error-390.png",
"sha256": "7bf3c42e0bcc7c76e0d5e7bdc73daa8a02bf2961146c8c550f1609fa0b759402"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\sse-raw-390.png",
"sha256": "8bdd5983369d1e2da56f66b81624c6d93025ab4649228b752e82cf725e81dc8c"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\probe-error-390.png",
"sha256": "7e8c19d768633df59cd301d5cc1051ad7652d89b59267be859fa4adbb73ffba2"
}
],
"errors": [],
"modelRequests": 3,
"realModelRequests": 0
}
Binary file not shown.

After

Width:  |  Height:  |  Size: 48 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 92 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 48 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 41 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 89 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 89 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 53 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 46 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 144 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 49 KiB

@@ -0,0 +1,57 @@
{
"success": true,
"appRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-win-x64-20261101140547938",
"checks": [
"native parameters persist after reload and the narrow form remains usable without calling a model",
"unsupported field is rejected visibly before any dispatch",
"actual HTTP error shows raw fields and original wait, keeps secrets hidden and downloads SHA-verified original at both widths",
"physical request ledger preserves the same error origin and offers the sealed actual request",
"Responses failed event retains original identity and reported use at both widths; partial tool is never dispatched or retried",
"explicit synthetic probe also shows the API diagnostic and labels its preview-only retention without inventing a Task artifact"
],
"screenshots": [
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\native-desktop.png",
"sha256": "7eb38b0ccfb87074d9d6011d1b869142b9d0e54ff81ecc951b38845219e7a80f"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\native-390.png",
"sha256": "bf0c6db94447284972e44f196b5136a33a41e71510deec9d5d09202e9290537e"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\http-error-desktop.png",
"sha256": "12ffc3b22da504d111919bf02bf287f9bb2ebca65b36117b843367e38345d9df"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\http-raw-desktop.png",
"sha256": "546c5382e99889e3b805add0bc97fe37442abbc13fd16577ad47a073d4f11ef0"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\http-error-390.png",
"sha256": "08f5ced640518714ae99504b58b4986cf1ac53f3a6fc09653e908c2fa59eb54c"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\ledger-390.png",
"sha256": "a6e15fe4b02b1263609bfaeed0dc809445a28b5902d16c5378af6a7c0768d078"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\sse-error-desktop.png",
"sha256": "f6d6c5e0804730dbab546c7ec57288ff32ba39c65b40c2251585ef7f4fd9d750"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\sse-error-390.png",
"sha256": "1dbb58980fba6d2282edf328c410d07799839deb49836aa5c8faee067c8dd0d6"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\sse-raw-390.png",
"sha256": "8bdd5983369d1e2da56f66b81624c6d93025ab4649228b752e82cf725e81dc8c"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\probe-error-390.png",
"sha256": "90a122e70ae15238e3acd85040b7346a4dadc5bd9c722f14608647ff5448b5f0"
}
],
"errors": [],
"modelRequests": 3,
"realModelRequests": 0
}
Binary file not shown.

After

Width:  |  Height:  |  Size: 48 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 93 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 48 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 149 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 116 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 149 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 116 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 41 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 44 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 101 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 100 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 42 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 48 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 42 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 89 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 52 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 46 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 144 KiB

@@ -0,0 +1,44 @@
{
"success": true,
"appRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench",
"checks": [
"native parameters persist after reload and the narrow form remains usable without calling a model",
"unsupported field is rejected visibly before any dispatch",
"actual HTTP error shows raw fields and original wait, keeps secrets hidden and downloads SHA-verified original at both widths",
"physical request ledger preserves the same error origin and offers the sealed actual request",
"Responses failed event retains original identity and reported use at both widths; partial tool is never dispatched or retried"
],
"screenshots": [
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-f256eff7-e02f-4186-a9e6-d0f879fe82fe\\native-desktop.png",
"sha256": "7eb38b0ccfb87074d9d6011d1b869142b9d0e54ff81ecc951b38845219e7a80f"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-f256eff7-e02f-4186-a9e6-d0f879fe82fe\\native-390.png",
"sha256": "bf0c6db94447284972e44f196b5136a33a41e71510deec9d5d09202e9290537e"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-f256eff7-e02f-4186-a9e6-d0f879fe82fe\\http-error-desktop.png",
"sha256": "1e5b870aa30622e7b40bea51a983f0199fe51d30e34fce2baec5b4e6917a6adc"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-f256eff7-e02f-4186-a9e6-d0f879fe82fe\\http-error-390.png",
"sha256": "d0b383a4f4d6716f65d92fa072cd546c7488fc371910b528006743478bbea5a2"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-f256eff7-e02f-4186-a9e6-d0f879fe82fe\\ledger-390.png",
"sha256": "203a5ce85361c5c444f62a43c0277e24525e1c6f4ca19808c18bf4c0cd161b22"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-f256eff7-e02f-4186-a9e6-d0f879fe82fe\\sse-error-desktop.png",
"sha256": "c849776a7a49b4fd2fc29b632ba4719d6004abe12676f8c3b1a39057195125e7"
},
{
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-f256eff7-e02f-4186-a9e6-d0f879fe82fe\\sse-error-390.png",
"sha256": "20d3270db03ac8edce6b6af2db48346f39150ab7c607a33d19936940081b352c"
}
],
"errors": [],
"modelRequests": 2,
"realModelRequests": 0
}
Binary file not shown.

After

Width:  |  Height:  |  Size: 47 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 89 KiB

@@ -0,0 +1 @@
{"success":true,"inputs":362,"compiled":22,"portableRoot":"D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-win-x64-20261101141644779"}
@@ -0,0 +1,16 @@
vite v7.3.7 building client environment for production...
transforming...
✓ 1736 modules transformed.
rendering chunks...
computing gzip size...
dist-client/.vite/manifest.json 0.55 kB │ gzip: 0.22 kB
dist-client/index.html 0.58 kB │ gzip: 0.40 kB
dist-client/mobile.html 0.82 kB │ gzip: 0.46 kB
dist-client/assets/mobile-CNDnI_ru.css 11.92 kB │ gzip: 3.12 kB
dist-client/assets/desktop-CeGvCGWP.css 149.17 kB │ gzip: 28.30 kB
dist-client/assets/mobile-DVWh54VU.js 24.31 kB │ gzip: 8.65 kB
dist-client/assets/x-Cc3T3VuL.js 225.51 kB │ gzip: 70.77 kB
dist-client/assets/desktop-BpRONdU9.js 332.48 kB │ gzip: 94.43 kB
✓ built in 1.22s
Mobile shell built: 3 allowed assets; API and project data are never cached.
Server bundle built.
@@ -0,0 +1,16 @@
vite v7.3.7 building client environment for production...
transforming...
✓ 1736 modules transformed.
rendering chunks...
computing gzip size...
dist-client/.vite/manifest.json 0.55 kB │ gzip: 0.22 kB
dist-client/index.html 0.58 kB │ gzip: 0.39 kB
dist-client/mobile.html 0.82 kB │ gzip: 0.46 kB
dist-client/assets/mobile-CNDnI_ru.css 11.92 kB │ gzip: 3.12 kB
dist-client/assets/desktop-CeGvCGWP.css 149.17 kB │ gzip: 28.30 kB
dist-client/assets/mobile-DVWh54VU.js 24.31 kB │ gzip: 8.65 kB
dist-client/assets/x-Cc3T3VuL.js 225.51 kB │ gzip: 70.77 kB
dist-client/assets/desktop-CvvqbCFw.js 332.08 kB │ gzip: 94.25 kB
✓ built in 1.17s
Mobile shell built: 3 allowed assets; API and project data are never cached.
Server bundle built.
@@ -0,0 +1,7 @@
PASS native parameters persist after reload and the narrow form remains usable without calling a model
PASS unsupported field is rejected visibly before any dispatch
PASS actual HTTP error shows raw fields and original wait, keeps secrets hidden and downloads SHA-verified original at both widths
PASS physical request ledger preserves the same error origin and offers the sealed actual request
PASS Responses failed event retains original identity and reported use at both widths; partial tool is never dispatched or retried
PASS explicit synthetic probe also shows the API diagnostic and labels its preview-only retention without inventing a Task artifact
{"success":true,"directory":"D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d","checks":6,"screenshots":10}
@@ -0,0 +1,60 @@
✔ HTTP quota retains native code/type/param/request id, bounded original and allowlisted headers, never repeats that connection (2437.6893ms)
✔ Retry-After above 90 seconds remains exact and a short local budget cannot trigger an early retry (1.2077ms)
✔ Retry-After date, expired date, huge value and invalid header retain explicit semantics (0.7535ms)
✔ unclassified 429 remains unspecified and every bounded actual retry retains its own original error (1809.9601ms)
✔ oversized HTTP error saves exactly the bounded prefix and labels its incomplete range (2023.3573ms)
✔ Chat SSE error preserves upstream fields, original failed event, partial tool and reported usage without executing or retrying (2273.9976ms)
✔ Responses failed SSE retains original event/response identity and usage (2112.0924ms)
✔ Responses incomplete SSE retains original event/response identity and usage (2243.0287ms)
✔ unsupported successful Responses output retains original JSON and usage with a local protocol origin (2230.7997ms)
✔ HTTP 200 native error is still an upstream failure rather than a fabricated successful answer (1.0275ms)
✔ quota failover uses another independent pool member and never replays a failed stream (2690.8242ms)
✔ native parameters are isolated by pool member, preserve Responses state and bind the original output reservation (0.4831ms)
✔ unsupported native fields and invalid values reject before dispatch instead of silently dropping or overriding constraints (0.4777ms)
✔ native settings for legacy and registered profiles persist without changing credentials or old records and reject stale edits (1253.3157ms)
✔ compiled native settings API saves/restores exact parameters and rejects reserved fields without mutating the profile (2376.7134ms)
✔ saved engine budgets verify all Chat/Responses route variants and reject a mismatched output allowance (1101.0856ms)
✔ owned CC Switch SQLite imports the selected TOML Responses wire_api and leaves unknown clients untouched (1132.6919ms)
✔ Responses executes one original tool through Runtime, carries encrypted Items, reports raw usage and seals exact wire bytes (2856.1773ms)
✔ mixed-protocol pool keeps both actual variants and changes route on 429 without re-executing completed tools (4046.8484ms)
▶ 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" (2055.2747ms)
✔ Reject {"body":{"id":"resp_owned","status":"failed","output":[{"type":"function_call","id":"fc_failed","cal (2433.8839ms)
✔ Reject {"body":{"id":"resp_owned","status":"completed","output":[{"type":"function_call","id":"fc_bad","cal (2285.8156ms)
✔ Reject {"body":{"id":"resp_owned","status":"completed","output":[{"type":"computer_call","id":"not-supporte (2036.6117ms)
✔ Responses incomplete, failed, missing call identity and unsupported hosted outputs never dispatch a tool (8813.8272ms)
▶ invalid JSON in a nonstream Chat or Responses tool group prevents every tool and pool replay
✔ chat-completions (2175.5782ms)
✔ responses (1643.874ms)
✔ invalid JSON in a nonstream Chat or Responses tool group prevents every tool and pool replay (3820.4375ms)
✔ wire adapter preserves tool IDs but never passes Responses-private Items to Chat or another endpoint (0.2013ms)
✔ connection JSON import accepts explicit Responses protocol and continues to reject unknown protocols without reading client sessions (885.7851ms)
✔ older configurations receive bounded wait defaults without changing the original object (1.0055ms)
✔ active stream survives sixty seconds; byte activity resets idle but cannot extend total deadline (0.8064ms)
✔ silent stream has its own accurate deadline; caller cancellation and cleanup preserve their cause (1.2455ms)
✔ waiting for headers is bounded independently of stream-idle time (0.224ms)
✔ wait settings persist, pin new runs and leave model pool, credentials and old run snapshots intact (1272.8715ms)
✔ direct code evidence remains selected while recent historical-index browsing is archived within the complete request budget (1396.7229ms)
✔ one long user turn becomes bounded complete groups with exact UTF-8 history and recent JSON arguments (2795.6134ms)
✔ tool definitions, exact system bytes, output reserve and remaining Run/Task budget are accounted before requests (637.6869ms)
✔ oversized required user instructions pause intact and can be explicitly resumed after increasing the window (1361.7653ms)
✔ all task instructions, constraints and acceptance remain in the runtime system instead of an eight-item or byte prefix (924.7642ms)
✔ a newest manageable multi-tool group stays exact even beyond the working target; oversized earlier groups are archived (777.1792ms)
✔ invalid orphan result is rejected before archive or network preparation (529.1ms)
✔ invalid missing result is rejected before archive or network preparation (577.0931ms)
✔ invalid wrong result is rejected before archive or network preparation (586.7706ms)
✔ invalid duplicate call is rejected before archive or network preparation (647.977ms)
✔ a group larger than the artifact limit uses verified Unicode-safe chunks without replacing tool JSON (1470.7051ms)
✔ changed or missing archived evidence stops reuse and is never silently recreated (805.1572ms)
✔ transaction failure prevents publishing a plan and any corresponding model request (773.6308ms)
✔ the Provider rejects body mutation after a sealed plan before calling HTTP (690.0836ms)
✔ HTTP failure and explicit same-Task resume reuse exact completed results and request hashes without repeating a command (5816.1929ms)
✔ a real process exit after plan persistence marks interruption and never issues an automatic model request (1278.9239ms)
ℹ tests 50
ℹ suites 0
ℹ pass 50
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 44053.6214
@@ -0,0 +1,114 @@
✔ the bundled local service enforces host, origin, token and JSON object contracts (2949.4169ms)
✔ an in-flight connection probe prevents configuration changes and another probe until its actual result returns (2892.5118ms)
✔ API review applies verified content, detects conflicts, exports sessions and keeps keys in memory (5546.3248ms)
✔ a real bundle restart preserves proposals, interrupts old approvals and requests fresh consent on resume (12349.391ms)
✔ runtime diagnostics and model-response validation use actual records and preserve the saved connection (2581.2984ms)
✔ persistent event replay is read-only and keeps run/artifact identity across reconnects (9970.8921ms)
✔ a question about the previous workbench error receives local error context and current change states (2336.1731ms)
✔ a transient model failure after an approved command retries only the model request (2994.2989ms)
✔ exhausted model retries preserve tool results, and explicit resume adds no new user turn or replay (4205.7639ms)
✔ stopping during model backoff cancels further requests and clears the running request state (2058.7019ms)
✔ HTTP authentication errors are classified, redacted and never automatically retried (21.3341ms)
✔ temporary HTTP service errors retry and recover with recorded attempt counts (13.0196ms)
✔ invalid and oversized protocol responses fail once with a useful response-format diagnosis (3.9587ms)
✔ HTTP rate limits expose their source and obey the service retry interval before recovery (1005.3494ms)
✔ Windows commands preserve Chinese through a PowerShell pipeline into Node (1113.0492ms)
✔ Windows shell syntax errors stay readable and native command exit codes are retained (1201.7401ms)
✔ review and completed states do not turn a generic continue click into a new model request (788.8897ms)
✔ old session data receives a truthful historical connection diagnosis without rewriting history (989.5805ms)
✔ model registry keeps original credentials/endpoint intact, rejects stale edits, and forgets only volatile keys on restart (1782.7768ms)
✔ config import previews exact revisions, keeps secrets volatile, syncs only when enabled and preserves manual edits on conflicts (981.9393ms)
✔ real HTTP pool fails over 429 to another model and never repeats completed tools; attempts seal actual request bytes and Run identity (4601.68ms)
✔ pool rotation and weights select members on new Runs; disabled and missing credentials are excluded without requests (5843.2078ms)
✔ 401 blocks the credential revision; updating the key restores availability without leaking error echoes (2539.5612ms)
✔ queued request cancellation and in-flight disable release only owned leases and do not dispatch another HTTP request (1772.6783ms)
✔ compiled model-center HTTP API persists public profiles, blocks forged origins, and local operations never invoke a model (2723.9885ms)
✔ model registry disk failures restore routing, pool profiles and credential references without partial updates (957.2737ms)
✔ model selection and saved configuration commit together and failures preserve the original mode and endpoint (1196.1936ms)
✔ configuration imports roll back the entire batch and retain one usable preview after validation or commit failure (1409.3272ms)
✔ HTTP quota retains native code/type/param/request id, bounded original and allowlisted headers, never repeats that connection (2948.9702ms)
✔ Retry-After above 90 seconds remains exact and a short local budget cannot trigger an early retry (1.3238ms)
✔ Retry-After date, expired date, huge value and invalid header retain explicit semantics (0.9156ms)
✔ unclassified 429 remains unspecified and every bounded actual retry retains its own original error (3195.1924ms)
✔ oversized HTTP error saves exactly the bounded prefix and labels its incomplete range (2698.3724ms)
✔ Chat SSE error preserves upstream fields, original failed event, partial tool and reported usage without executing or retrying (2487.1155ms)
✔ Responses failed SSE retains original event/response identity and usage (2678.1828ms)
✔ Responses incomplete SSE retains original event/response identity and usage (3326.3782ms)
✔ unsupported successful Responses output retains original JSON and usage with a local protocol origin (3454.4363ms)
✔ HTTP 200 native error is still an upstream failure rather than a fabricated successful answer (0.791ms)
✔ JSON-escaped credential echoes are redacted in the original preview and capture as well as parsed fields (0.823ms)
✔ quota failover uses another independent pool member and never replays a failed stream (3694.3879ms)
✔ native parameters are isolated by pool member, preserve Responses state and bind the original output reservation (0.3413ms)
✔ unsupported native fields and invalid values reject before dispatch instead of silently dropping or overriding constraints (0.2947ms)
✔ native settings for legacy and registered profiles persist without changing credentials or old records and reject stale edits (1818.9464ms)
✔ compiled native settings API saves/restores exact parameters and rejects reserved fields without mutating the profile (3112.1406ms)
✔ legacy connection test sends its saved native fields and selected output limit field without changing the configuration (3782.7477ms)
✔ saved engine budgets verify all Chat/Responses route variants and reject a mismatched output allowance (1744.7918ms)
✔ owned CC Switch SQLite imports the selected TOML Responses wire_api and leaves unknown clients untouched (1738.9581ms)
✔ Responses executes one original tool through Runtime, carries encrypted Items, reports raw usage and seals exact wire bytes (3938.1637ms)
✔ mixed-protocol pool keeps both actual variants and changes route on 429 without re-executing completed tools (5928.3399ms)
▶ 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" (3219.1706ms)
✔ Reject {"body":{"id":"resp_owned","status":"failed","output":[{"type":"function_call","id":"fc_failed","cal (2782.9028ms)
✔ Reject {"body":{"id":"resp_owned","status":"completed","output":[{"type":"function_call","id":"fc_bad","cal (3311.2631ms)
✔ Reject {"body":{"id":"resp_owned","status":"completed","output":[{"type":"computer_call","id":"not-supporte (1711.9685ms)
✔ Responses incomplete, failed, missing call identity and unsupported hosted outputs never dispatch a tool (11026.1613ms)
▶ invalid JSON in a nonstream Chat or Responses tool group prevents every tool and pool replay
✔ chat-completions (2973.3093ms)
✔ responses (3122.6027ms)
✔ invalid JSON in a nonstream Chat or Responses tool group prevents every tool and pool replay (6096.3845ms)
✔ wire adapter preserves tool IDs but never passes Responses-private Items to Chat or another endpoint (0.1997ms)
✔ connection JSON import accepts explicit Responses protocol and continues to reject unknown protocols without reading client sessions (829.1866ms)
✔ HTTP download preserves original malformed SSE bytes and artifact SHA instead of decoding invalid UTF8 (3510.7978ms)
✔ terminated SSE adapts empty optional tool arguments, retains the exact wire bytes and executes the validated batch once (5124.9382ms)
✔ reported SSE usage survives rejected required arguments "" without dispatching any tool (3214.5679ms)
✔ reported SSE usage survives rejected required arguments "{\"path\":" without dispatching any tool (2732.8113ms)
✔ reported SSE usage survives rejected required arguments "null" without dispatching any tool (2480.8519ms)
✔ a stream without DONE retains reported usage but cannot adapt or execute even an optional tool (3011.9954ms)
✔ a timeout after reasoning-only fragments creates one unknown receipt and never invents reported tokens (4326.249ms)
✔ real Chat SSE carries UTF8 and split tool JSON; Runtime executes only after DONE, stores partial progress and original usage (5749.6896ms)
✔ Responses typed SSE validates original item identity, terminal data, raw usage and tool result pair (5671.6348ms)
✔ incomplete network stream retains unknown attempt and never automatically switches pool or dispatches fragment tools (3099.9542ms)
✔ stopping an active HTTP stream closes only original request and keeps partial text without executing tools (2828.8084ms)
✔ explicit streaming probe uses actual SSE despite unknown declaration and binds its own original attempt (1339.6748ms)
✔ coalesced initial SSE events publish the latest partial state even when the upstream stays idle (2510.4091ms)
✔ stream timeout closes the original HTTP response, preserves redacted capture and never retries (1683.4218ms)
✔ stream idle timeout preserves incomplete tool arguments, records unknown use and never fails over (2137.8733ms)
✔ SSE accepts CR-only final event boundaries and retains exact invalid UTF8 bytes without dispatch (0.7074ms)
▶ stream parser rejects truncation, finish length, malformed JSON, duplicate sequence, changed final item and oversized data
✔ reject 0 (0.2013ms)
✔ reject 1 (0.1345ms)
✔ reject 2 (0.1293ms)
✔ reject 3 (4.8833ms)
✔ reject 4 (0.1962ms)
✔ reject 5 (0.1977ms)
✔ stream parser rejects truncation, finish length, malformed JSON, duplicate sequence, changed final item and oversized data (6.4022ms)
✔ older configurations receive bounded wait defaults without changing the original object (0.9937ms)
✔ active stream survives sixty seconds; byte activity resets idle but cannot extend total deadline (0.7834ms)
✔ silent stream has its own accurate deadline; caller cancellation and cleanup preserve their cause (1.4035ms)
✔ waiting for headers is bounded independently of stream-idle time (0.2343ms)
✔ wait settings persist, pin new runs and leave model pool, credentials and old run snapshots intact (1925.5894ms)
✔ direct code evidence remains selected while recent historical-index browsing is archived within the complete request budget (2710.3659ms)
✔ one long user turn becomes bounded complete groups with exact UTF-8 history and recent JSON arguments (7119.847ms)
✔ tool definitions, exact system bytes, output reserve and remaining Run/Task budget are accounted before requests (1105.2088ms)
✔ oversized required user instructions pause intact and can be explicitly resumed after increasing the window (2178.405ms)
✔ all task instructions, constraints and acceptance remain in the runtime system instead of an eight-item or byte prefix (2167.4349ms)
✔ a newest manageable multi-tool group stays exact even beyond the working target; oversized earlier groups are archived (2301.5279ms)
✔ invalid orphan result is rejected before archive or network preparation (1526.6228ms)
✔ invalid missing result is rejected before archive or network preparation (1190.5211ms)
✔ invalid wrong result is rejected before archive or network preparation (1431.5339ms)
✔ invalid duplicate call is rejected before archive or network preparation (1398.6444ms)
✔ a group larger than the artifact limit uses verified Unicode-safe chunks without replacing tool JSON (2598.6125ms)
✔ changed or missing archived evidence stops reuse and is never silently recreated (2088.2513ms)
✔ transaction failure prevents publishing a plan and any corresponding model request (1608.6043ms)
✔ the Provider rejects body mutation after a sealed plan before calling HTTP (1549.1033ms)
✔ HTTP failure and explicit same-Task resume reuse exact completed results and request hashes without repeating a command (13459.2583ms)
✔ a real process exit after plan persistence marks interruption and never issues an automatic model request (2404.0743ms)
ℹ tests 103
ℹ suites 0
ℹ pass 103
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 121226.9117
@@ -0,0 +1 @@
{"success":true,"inputs":362,"compiled":22,"portableRoot":"D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-win-x64-20261101140547938"}
@@ -0,0 +1,7 @@
PASS native parameters persist after reload and the narrow form remains usable without calling a model
PASS unsupported field is rejected visibly before any dispatch
PASS actual HTTP error shows raw fields and original wait, keeps secrets hidden and downloads SHA-verified original at both widths
PASS physical request ledger preserves the same error origin and offers the sealed actual request
PASS Responses failed event retains original identity and reported use at both widths; partial tool is never dispatched or retried
PASS explicit synthetic probe also shows the API diagnostic and labels its preview-only retention without inventing a Task artifact
{"success":true,"directory":"D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860","checks":6,"screenshots":10}
@@ -0,0 +1,11 @@
{"success":false,"error":"Error: This migration accepts a supported existing Workbench.\n at upgradeRunning (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/scripts/upgrade-running.mjs:67:324)\n at process.processTicksAndRejections (node:internal/process/task_queues:104:5)\n at async file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/.qa/model-fidelity-owned-upgrade.mjs:17:20"}
node:internal/modules/run_main:107
triggerUncaughtException(
^
Error: This migration accepts a supported existing Workbench.
at upgradeRunning (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/scripts/upgrade-running.mjs:67:324)
at process.processTicksAndRejections (node:internal/process/task_queues:104:5)
at async file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/.qa/model-fidelity-owned-upgrade.mjs:17:20
Node.js v24.22.1
@@ -0,0 +1,70 @@
{
"portableRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-win-x64-20261101140547938",
"sourceRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-source-20261101140547938",
"portable": {
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-win-x64-20261101140547938.zip",
"bytes": 629123043,
"sha256": "419dfe81ab32e3dc0abc23f2714e2f603bc06c4fb1941f6db4dca050204e8b19"
},
"source": {
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-source-20261101140547938.zip",
"bytes": 629305353,
"sha256": "8cfda0726019d9bb119cc4b9be0aff7a4cb4c76d61106791060315ac920ad56a"
},
"installer": {
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-win-x64-20261101140547938-Setup.exe",
"bytes": 588262697,
"sha256": "5cb96077abc5b6a4339475228804761df65907004c1904f221827f5c71623f46",
"compiler": {
"version": "6.3.1",
"sha256": "3b22d36fc7e7c421d8caf780008ecdb42b14b3f0371b9bafe3cad64e6648d9cf"
},
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\installer-build-18e26917-bee3-4b41-bdc4-09f7a01ed915\\installer.iss",
"version": "1.20.3",
"candidateDirectory": "Agent-Workbench-1.20.3-win-x64-20261101140547938",
"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": 19638,
"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 @@
{"success":false,"phase":"activate","output":".qa/model-fidelity-formal-activate.json","error":"AssertionError [ERR_ASSERTION]: sessionsSha256\n+ actual - expected\n\n+ '11c2bca842fa257402b23224957a23d11441b86aa2b94e6a4344a0e8960c579e'\n- 'cdecf3e1f8b3ba36d168e17394ac509ec8394b8adc7748e43b7e21bf16482168'\n\n at compareOriginalState (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/.qa/desktop-f-state.mjs:47:116)\n at file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/.qa/model-fidelity-formal.mjs:69:69\n at process.processTicksAndRejections (node:internal/process/task_queues:104:5)"}
@@ -0,0 +1 @@
{"success":true,"phase":"staged","output":".qa/model-fidelity-formal-stage.json"}
@@ -0,0 +1,21 @@
node:internal/modules/run_main:107
triggerUncaughtException(
^
AssertionError [ERR_ASSERTION]: Expected values to be strictly equal:
+ actual - expected
+ 'completed'
- 'failed'
at file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/.qa/model-fidelity-formal-ui.mjs:15:8
at process.processTicksAndRejections (node:internal/process/task_queues:104:5) {
generatedMessage: true,
code: 'ERR_ASSERTION',
actual: 'completed',
expected: 'failed',
operator: 'strictEqual',
diff: 'simple'
}
Node.js v24.22.1
@@ -0,0 +1 @@
{"success":false,"results":1,"screenshots":2,"error":"Error: expect(locator).toBeVisible() failed\n\nLocator: getByText('创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D', { exact: true }).first()\nExpected: visible\nReceived: hidden\nTimeout: 5000ms\n\nCall log:\n - Expect \"toBeVisible\" getByText('创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D', { exact: true }).first() with timeout 5000ms\n - waiting for getByText('创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D', { exact: true }).first()\n 14 × locator resolved to <span>创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D</span>\n - unexpected value \"hidden\"\n\n at captureRawStack (D:\\Ai-project\\REA逆向\\Agent-Workbench\\node_modules\\playwright-core\\lib\\coreBundle.js:9282:17)\n at callMatcherAsStep (D:\\Ai-project\\REA逆向\\Agent-Workbench\\node_modules\\playwright\\lib\\matchers\\expect.js:13316:57)\n at Object.toBeVisible (D:\\Ai-project\\REA逆向\\Agent-Workbench\\node_modules\\playwright\\lib\\matchers\\expect.js:13308:23)\n at inspect (D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui.mjs:30:81)\n at async file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/.qa/model-fidelity-formal-ui.mjs:52:5"}
@@ -0,0 +1 @@
{"success":true,"results":4,"screenshots":8}
@@ -0,0 +1 @@
{"success":true,"phase":"upgraded","output":".qa/model-fidelity-formal-upgrade.json"}
@@ -0,0 +1 @@
{"success":true,"retained":{"originalRows":"exact except modelAttempts appended two owned user attempts and events sequence advanced","originalExports":"all 31 exact","originalArtifactBytes":"all original exact; nine new same-session artifacts retained","originalModelCenter":"all profiles/routes/credentials/original attempts exact; two user attempts and their connection health retained","stableCurrentRowsExportsArtifactsAndModelCenter":{"originalRows":"exact","originalExports":"exact","originalArtifactBytes":"exact","originalModelCenter":"exact"}},"additions":{"sessions":1,"records":16,"events":57,"outbox":57,"artifacts":9,"attempts":2},"counts":{"sessions":32,"records":3239,"events":6229}}
@@ -0,0 +1,718 @@
✔ the bundled local service enforces host, origin, token and JSON object contracts (2135.8701ms)
✔ an in-flight connection probe prevents configuration changes and another probe until its actual result returns (2490.8502ms)
✔ API review applies verified content, detects conflicts, exports sessions and keeps keys in memory (3866.4687ms)
✔ a real bundle restart preserves proposals, interrupts old approvals and requests fresh consent on resume (9264.4658ms)
✔ runtime diagnostics and model-response validation use actual records and preserve the saved connection (2272.882ms)
✔ persistent event replay is read-only and keeps run/artifact identity across reconnects (5774.1107ms)
✔ real browser HTTP plans, original field pages and PNG bytes preserve ownership and survive host restart (7479.0599ms)
✔ self harness browser tools require approval and feed verified actual page fields back to the original Task (11215.8639ms)
✔ browser host recovery retains unknown root outcome, blocks replay, and requires an explicit owned acknowledgement (4131.4615ms)
✔ late HTTP stop of a finished browser job cannot interrupt a later job in the same Task (7781.9882ms)
✔ actual failed page blocks source-passed delivery; a new passing page permits a bound commit and preserves both original records (19288.8061ms)
✔ a build-only change after Git review invalidates the actual browser evidence and leaves HEAD and the protected index intact (10355.4518ms)
✔ actual DOM controls include validated unique CSS selectors across repeated parent buttons and escaped labels (7318.8067ms)
✔ browser field coverage distinguishes sections and Run windows while retaining original source identity (0.7568ms)
✔ repeated original browser pages pause a live Task without re-opening the browser or breaking tool pairing (4808.5154ms)
✔ browser plan rejects outside builds, invalid interaction scope and unbounded steps before execution (0.9065ms)
✔ real owned Edge interactions, assertions, raw PNG and stopped process survive exact result reuse and host reopen (3883.7258ms)
✔ failed real page assertion preserves actual value and forbids silent retry of the original plan (8513.1066ms)
✔ build/source mutation and sealed step tampering reject before browser dispatch; earlier task cannot execute plan (2038.3713ms)
✔ outside network never reaches another local service, and original JavaScript/console failures remain failures (3126.2476ms)
✔ stopping a real in-flight browser preserves finished steps, closes its owned root and never replays (3439.3012ms)
✔ 原PNG损坏使原结果校验失败,不能以已有通过状态复用 (3269.8709ms)
✔ a budget pause persists concrete text/tool/history/image breakdown and never dispatches an oversized request (2031.2413ms)
✔ explicit frozen preview resumes the same Task in a new Run, preserves completed tools and does not change global config or selected model (2467.1438ms)
✔ insufficient budgets, silently clamped capacity and arbitrary config fields reject without a physical request or mutating task budget (2532.9972ms)
✔ changed session, changed registry, expiry and altered frozen preview all require a new explicit preview (2806.0112ms)
✔ preview preserves unknown invocation status and increasing Task allowance is committed only with confirmed new Run (3155.5519ms)
✔ batch APIs enforce ownership and review fingerprints, apply once, and persist the completed batch across restart (5329.1084ms)
✔ delivery reports bind actual batch identity and reject corrupted snapshot evidence without changing project files (4476.122ms)
✔ a compatible model can seal proposals through the real HTTP loop while application stays behind user review (7139.4349ms)
✔ a step-limited sealed batch applies and continues the same Task into actual checks without replaying saved tools (13317.131ms)
✔ applied batches cannot turn cancellation, other budgets, restart or unknown commands into a review continuation (2044.2172ms)
✔ actual process exit at file-batch prepared preserves durable identity and never replays writes (1153.2736ms)
✔ actual process exit at file-batch application_started preserves durable identity and never replays writes (1446.7508ms)
✔ actual process exit at file-batch write_intent preserves durable identity and never replays writes (1299.269ms)
✔ actual process exit at file-batch staged preserves durable identity and never replays writes (2496.8656ms)
✔ actual process exit at file-batch replaced preserves durable identity and never replays writes (1849.1501ms)
✔ actual process exit at file-batch file_recorded preserves durable identity and never replays writes (2125.7141ms)
✔ actual process exit at file-batch finished preserves durable identity and never replays writes (2282.7416ms)
✔ actual process exit at file-batch rollback_started preserves durable identity and never replays writes (2686.2087ms)
✔ actual process exit at file-batch restore_intent preserves durable identity and never replays writes (2971.0743ms)
✔ actual process exit at file-batch restore_staged preserves durable identity and never replays writes (3001.0747ms)
✔ actual process exit at file-batch restored preserves durable identity and never replays writes (2717.3227ms)
✔ actual process exit at file-batch restore_recorded preserves durable identity and never replays writes (2593.2422ms)
✔ actual process exit at file-batch rollback_finished preserves durable identity and never replays writes (2879.0374ms)
warning: in the working copy of 'README.md', LF will be replaced by CRLF the next time Git touches it
warning: in the working copy of 'src/two.mjs', LF will be replaced by CRLF the next time Git touches it
✔ sealed batches preserve bytes until review, then apply and explicitly restore BOM/CRLF/new files (2571.63ms)
Switched to a new branch 'owner-branch'
✔ a conflict in the last file or Git branch prevents writing every file in the batch (3247.413ms)
✔ duplicate canonical paths, ancestor collisions, links, dependencies and Windows aliases never form a batch (807.0127ms)
✔ fingerprints, changed proposals and corrupt snapshot artifacts cannot write or restore any target (2214.8295ms)
✔ a failure after replacement restores only batch bytes and never retries an old application (1701.8152ms)
✔ rollback preserves owner edits, blocks other project writers, and explicit acknowledgement releases recovery (1554.5079ms)
✔ an unattempted owner-created file identical to a suggestion remains external and is never removed (1600.1951ms)
✔ active batch requests are serialized and cancellation leaves a verifiable ended batch (1592.7204ms)
✔ batch application and rollback preserve unrelated Git index and worktree bytes (3310.9221ms)
✔ recovery reuses exact staged bytes but retains a conflicting temporary file for owner inspection (3242.4297ms)
✔ one Task can seal pending proposals from earlier Runs while snapshots retain each source Run (2092.947ms)
✔ engineering inspection preserves a dirty Git baseline and treats project rules as data (3141.8175ms)
✔ real npm tests and build produce counted results, log artifacts and persistent verification (2676.1304ms)
✔ an actual failing test retains the nonzero exit code and failed count (2264.2343ms)
✔ unchanged automatic checks return original failure evidence, while explicit user reruns and changed source still execute (4549.1354ms)
✔ agent loops cannot hide identical failing checks behind new plan IDs or unrelated reads (8513.8971ms)
✔ zero tests, skipped tests and unrecognized stdout cannot become verification passes (8185.8321ms)
✔ a successful build with no selected tests does not verify the task (1659.6287ms)
✔ stale check plans and forged script names execute no project commands (1116.3956ms)
✔ applying a later proposal invalidates the prior passed engineering snapshot (3356.8867ms)
✔ stopping an actual check terminates its child process and blocks concurrent jobs (2517.6485ms)
✔ check timeout records a failure with its actual termination result (2461.3459ms)
✔ source edits during a successful check make the result stale (3397.5359ms)
✔ bounded large logs retain the test footer and readable artifact instead of inventing counts (2314.3067ms)
✔ model-selected checks require a run-bound decision and denial creates no check job (2455.1095ms)
✔ check APIs expose owned plans and jobs and preserve the authoritative results across restart (7278.2346ms)
✔ reviewed changes continue the original task without reopening a generic completed run (2845.7217ms)
✔ concurrent review-continuation API requests accept one run and retain the goal after restart (9020.4263ms)
✔ the model reads owned failure logs without running the check again (3284.7165ms)
✔ process exit at check queued preserves results and never replays commands (1330.5533ms)
✔ process exit at check result_committed preserves results and never replays commands (2070.6431ms)
✔ model code tools keep Task/Run ownership, index evidence and exact proposals across storage recovery (5610.4894ms)
✔ bundled code APIs isolate project results, validate queries and create review-only patches with conflicts (3386.2914ms)
✔ code tool arguments reject malformed ownership-independent inputs before side effects (0.3251ms)
✔ observe-only device authority cannot read, configure or clear project parsing caches (1214.3094ms)
✔ pairing is short lived, single-use, generation-bound and validates exact permissions (662.5548ms)
✔ device transaction failure neither consumes the ticket nor leaves a credential (608.8573ms)
✔ devices have project and action boundaries; rotation, expiration, revocation and host changes reject old authority (708.3191ms)
✔ persistent registry stores hashes and a bounded audit, never raw credentials (987.1752ms)
✔ real TLS verifies IP SAN, CA chain, encrypted restart and public fingerprint (812.3715ms)
✔ TLS sealing failure leaves listener closed and no plaintext or successful TLS registration (467.5834ms)
✔ HTTP pairing is concurrent single-use; secure cookie and strict origin/host checks (842.4674ms)
✔ gateway never exposes management routes, arbitrary assets or unpaired project data (914.0171ms)
✔ authorized project events paginate from original seq without leaking other project events or payloads (1046.2245ms)
✔ remote task projection excludes raw model history, host configuration, tool input and credentials (1003.0778ms)
✔ permission checks apply to direct remote mutation, not only UI controls (749.2914ms)
✔ approval fingerprint and exact task/revision are bound; duplicate requests never execute twice (751.4223ms)
✔ unknown original operation is recorded and never replayed after response failure (723.6303ms)
✔ revocation and gateway closure reject old access without invoking runtime stop (981.5965ms)
✔ rate limits reject pairing brute force and do not persist candidate secrets (850.8244ms)
✔ real Windows CurrentUser DPAPI persists only a sealed CA and reopens it (1758.9845ms)
✔ mobile approval executes the original Runtime command exactly once and rejects the later desktop decision (2573.6869ms)
✔ desktop rejection wins a concurrent mobile request and original command remains unexecuted (1578.6577ms)
✔ phone stop waits for the original process exit and prevents later heartbeats and command replay (2286.2328ms)
✔ Run replacement during parent-stop wait is rejected rather than stopping the new Run (539.7268ms)
✔ diff pages preserve source SHA; report reads verify blobs and exclude unrelated artifact kinds (884.2056ms)
✔ hot upgrade refuses an active gateway and restores closed gateway ownership on failure (1017.1472ms)
✔ damaged device permissions or expiry fail closed and never overwrite the original registry (925.8584ms)
✔ device revocation destroys its actual authenticated TLS connection (1070.6881ms)
✔ companion observes only current Task screenshots, verifies original PNG bytes and loses access immediately on revoke (1141.879ms)
✔ actual config HTTP failure preserves the committed selection, config and volatile credential before a successful retry (856.0603ms)
✔ controlled installation pins the native kernel and bridge independently of the existing connection-probe identity (890.9011ms)
✔ controlled profiles disable every native project/terminal provider and native model/retry route using the actual include patch contract (0.5464ms)
✔ owned tool catalogues seal full schemas canonically and reject duplicate or malformed declarations (4.9129ms)
✔ bridge authentication, current Run, process, SDK session and browser-origin checks reject before dispatch (837.7027ms)
✔ bridge duplicates, unknown tools, malformed arguments and source changes are refused before a second side effect (720.7667ms)
✔ bridge model requests bind the exact catalogue, route and output bound and preserve raw request/result identity (857.2781ms)
✔ stopped and oversized bridge requests never reach a local model or tool callback (703.1826ms)
✔ a Windows virtual data-directory name binds new storage and credentials to its actual physical root (1822.3752ms)
✔ physical data-root adoption still rejects explicit junctions at the directory or any ancestor (471.5053ms)
✔ actual Windows CurrentUser DPAPI roundtrips owned UTF-8 and rejects damaged bytes and wrong entropy (1487.3432ms)
✔ only explicitly supplied keys persist; new provider restores original revisions and public views contain no key (2497.5823ms)
✔ registry and cipher commit failure restores original routing, keys, block and every database row (1920.3182ms)
✔ a two-key import with a protection failure rolls back all registry and cipher entries and retains its preview (2301.4798ms)
✔ connection revision changes retain the original saved block and report a source mismatch on restart (2380.309ms)
✔ disable skips real restart loading, explicit enable restores, delete preserves memory and forget removes both (2491.2257ms)
✔ corrupt saved block and another SID do not load; failures are visible without returning ciphertext or plaintext (2588.241ms)
✔ legacy persistence requires a new explicit key and binds config and credential revisions; stale requests cannot save (2653.0914ms)
✔ owned plaintext keys never reach SQLite pages, WAL, config or public execution exports (1825.5967ms)
✔ actual process exit at Git queued never replays a commit and can reconcile the durable result (10487.3896ms)
✔ actual process exit at Git preparing never replays a commit and can reconcile the durable result (10500.414ms)
✔ actual process exit at Git index_locked never replays a commit and can reconcile the durable result (11698.8076ms)
✔ actual process exit at Git prepared never replays a commit and can reconcile the durable result (14087.1733ms)
✔ actual process exit at Git ref_prepared never replays a commit and can reconcile the durable result (14846.5685ms)
✔ actual process exit at Git publishing never replays a commit and can reconcile the durable result (14931.2276ms)
✔ actual process exit at Git committed never replays a commit and can reconcile the durable result (13457.0895ms)
✔ actual process exit at Git index_published never replays a commit and can reconcile the durable result (13027.9938ms)
✔ actual process exit at Git finished never replays a commit and can reconcile the durable result (11870.1597ms)
✔ an unknown Git result cannot overwrite an externally changed index or delete an unrelated lock (11820.1766ms)
✔ known publication with a later source edit retains the commit and marks the delivery for inspection (11525.688ms)
✔ CRLF normalization, multiple applied revisions and unrelated special index flags preserve exact bytes and ownership (12921.4106ms)
✔ adding new durable record kinds to an existing hot Store preserves all existing Maps and records (9773.4956ms)
✔ an unrelated owner deletion is fingerprinted and preserved while linked task paths are rejected (13025.6977ms)
✔ actual failed tests become a reviewed local commit and evidence report while preserving unrelated index and files (10863.2192ms)
✔ selected preexisting unstaged, staged and untracked owner edits cannot enter a task commit (18002.3573ms)
Switched to a new branch 'owner-branch'
✔ source, unrelated stage and branch changes after review invalidate a commit plan without changing HEAD (22812.0888ms)
Switched to a new branch 'delivery-branch'
✔ a later Git metadata change keeps matching source evidence valid but a real source edit requires checks (11304.2925ms)
✔ a corrupt passed-check log and a pending proposal block Git delivery instead of inventing evidence (7662.8351ms)
HEAD is now at e0b4ee0 Sealed delivery fixture baseline
✔ Git filters, merge state, detached HEAD and old task ownership are explicitly rejected (28034.9817ms)
✔ local snapshot commits disable repository hooks and signing and isolate the Git child environment (11907.6824ms)
✔ stopping before publication cancels safely; stopping after publication preserves the actual completed commit (19676.2168ms)
✔ delivery API confirms exactly one owned plan, serves verified reports and preserves results across restart (14812.2491ms)
✔ the model can prepare an owned Git review and evidence report while publishing remains a separate user action (10797.8913ms)
✔ HTTP Git review pairs deferred calls, survives reload and never requests more work until the saved preview is resolved (8319.1015ms)
✔ a reviewed Git preview can be revised or withdrawn without publishing or losing the current Task (7633.6892ms)
✔ actual Electron debug argument ordering retains exactly the compiled entry and rejects extra or repeated actions (1.0943ms)
✔ native bridge refuses other origins, subframes, other windows and unsupported inputs before connecting (613.1173ms)
✔ file reveal binds the current project, task, path, UTF-8 bytes and original SHA without executing the file (538.0116ms)
✔ file changed during asynchronous task recheck is not revealed as the old bytes (645.6827ms)
✔ desktop browser profile and aliases stay outside project inventory and file access (578.1162ms)
✔ native picker cancel and origin change after dialog publish no project registration (755.4598ms)
✔ pending native action is not dispatched twice, and current task review returns its original seq (624.7792ms)
✔ native stop cancellation and changed original host never stop a replacement process (791.6811ms)
✔ window register rejects host PID, stale installation, extra fields and missing creation identity (0.317ms)
✔ a rebuilt source archive binds actual generated client assets and retains its original running identity after disk changes (1571.8171ms)
✔ host challenge proof rejects a stale challenge, changed identity and changed signature (0.7297ms)
✔ stop confirmation binds the original instance, creation and loaded runtime (0.129ms)
✔ state paths reject escaping, linked ancestors, ambiguous hardlinks and damaged JSON (768.8181ms)
✔ failed registration publishes no initialized memory and can complete after its actual inputs are supplied (2405.2038ms)
✔ host shutdown cancels an un-signalled HTTP model request and refuses later requests until recovered (1.3817ms)
✔ host shutdown cancels model-pool acquisition and releases queue ownership without inventing attempts (682.2098ms)
✔ full candidate validation returns the exact loaded-input host identity and rejects undeclared or altered files (1307.9091ms)
✔ installation requires separate absolute program/data directories and a candidate in its own versions namespace (855.4176ms)
✔ staging retains the original data and pointer without querying or restarting a host (1858.7072ms)
✔ a post-pointer failure restores every original pointer byte and persistent row; a first-install failure removes only its published pointer (4444.2533ms)
✔ changed active manifest and concurrent pointer changes are rejected without overwriting the unknown record (2071.5227ms)
✔ an unregistered non-HTTP listener blocks preflight even when service-info is absent (1097.8719ms)
✔ uninstall archives its original pointer and preserves every data row/artifact instead of recursively deleting data (3551.335ms)
✔ preflight rejects an unrelated program root or changed native bootstrap while preserving its original bytes (1009.2965ms)
✔ a changed owned bootstrap follows its candidate and a post-pointer fault restores both original bytes (5330.3064ms)
✔ an external bootstrap change after preflight is preserved and never overwritten as an update or rollback (3504.6546ms)
✔ owned startup retries only read-only observation after a slow health response and attests the same original child (4189.6324ms)
✔ unverified startup expires with its original live child and never treats timeout or an unrelated health PID as proof (5192.3334ms)
✔ a question about the previous workbench error receives local error context and current change states (1757.752ms)
✔ a transient model failure after an approved command retries only the model request (2524.4187ms)
✔ exhausted model retries preserve tool results, and explicit resume adds no new user turn or replay (2925.0855ms)
✔ stopping during model backoff cancels further requests and clears the running request state (1006.782ms)
✔ HTTP authentication errors are classified, redacted and never automatically retried (21.5568ms)
✔ temporary HTTP service errors retry and recover with recorded attempt counts (20.0916ms)
✔ invalid and oversized protocol responses fail once with a useful response-format diagnosis (3.957ms)
✔ HTTP rate limits expose their source and obey the service retry interval before recovery (1011.6952ms)
✔ Windows commands preserve Chinese through a PowerShell pipeline into Node (736.7706ms)
✔ Windows shell syntax errors stay readable and native command exit codes are retained (850.8498ms)
✔ review and completed states do not turn a generic continue click into a new model request (513.3558ms)
✔ old session data receives a truthful historical connection diagnosis without rewriting history (704.2704ms)
✔ engine HTTP checks bind the detected installation and current task, page original bytes and preserve them across host restart (4945.5123ms)
✔ HTTP stop seals a pending SDK call, rejects automatic replay and keeps failures attached to the original session (3763.7387ms)
✔ controlled engine HTTP freezes one child, pages original owned input/report, rejects foreign reads and prevents changed-source reuse (12751.6464ms)
✔ native text and call IDs preserve original result pairing while unsupported content fails closed (0.7719ms)
✔ one engine child freezes real installation identity before creating tasks and rejects a changed source (7409.8399ms)
✔ owned native tools return saved results, use real approvals/checks and account each actual request once across new engine Runs (25797.7235ms)
✔ forged native scope cannot change a parent file and a final completion claim cannot replace worker proof (12438.6545ms)
✔ duplicate native IDs are rejected without replaying a saved tool (10939.5564ms)
✔ native input exceeding the context budget is retained and sends no model request (11602.5458ms)
✔ stop retains a late reported model usage once and executes no tool from that output (14235.9999ms)
✔ a source change after the original model intent stops before native tools execute (18226.8614ms)
✔ an engine source change while approval is pending invalidates the old permission before any check starts (25390.5438ms)
✔ missing or noncompleted turn end cannot be adopted even with an owned receipt and assistant text (0.6604ms)
✔ closed root with an original live descendant remains unknown and retains its process tree (0.2949ms)
✔ host recovery reserves original pending attempts exactly once and explicit review preserves unknown outcomes (28831.6282ms)
✔ corrupted model reservation cannot reduce recovered budget or permit a new engine request (15748.4958ms)
✔ engine detection seals the actual CLI, SDK, profile and Node and rejects mixed versions without launch (1494.3789ms)
✔ owned engine probe waits for real ready identity, verifies six native responses and saves immutable task evidence (1888.8926ms)
✔ an invalid startup nonce is rejected before initialize and preserves the process close and partial report (1480.8063ms)
✔ SDK identity and session directory mismatches retain raw responses and never become verified (4065.038ms)
✔ cancel during a pending initialize closes only the owned process without sending an unsupported MCP cancellation (1336.4129ms)
✔ cancel before readiness sends no protocol request and a user retry creates a different process and record (2004.9254ms)
✔ source or scratch changes and unexpected model traffic fail after preserving the actual observed result (4320.7941ms)
✔ unowned idle notifications stay quarantined and server requests receive a rejection without granting operations (3119.3103ms)
✔ engine probe environment excludes parent secrets and NODE_OPTIONS while preserving operating system paths (1143.2161ms)
✔ foreign sessions and altered reports or task bindings cannot reuse verified engine evidence (1245.8913ms)
✔ restart never replays an active engine check or stops a historical PID, and keeps prior task state (1674.0314ms)
✔ model argument validation and runtime activity do not grant external engine prompt or project execution (641.7865ms)
✔ a follow-up question receives the actual failed engine check from its original Task and keeps its report unchanged (2467.7849ms)
✔ evaluation denominator distinguishes attempted unsealed candidates from unexecuted samples (0.7584ms)
✔ evaluation summary retains sealed outcomes and stopped or failed original attempts (0.4654ms)
✔ evaluation contract fixes bounded samples, source paths, checks and aggregate budget (572.1567ms)
✔ same-source workspaces seal real commands before execution and preserve original staged and unstaged bytes (13720.9835ms)
✔ actual original failures and candidate success are both preserved, then adoption requires parent review, checks and Git (63568.2878ms)
✔ candidate cannot shrink tests and changing original test bytes rejects execution before any model request (31974.726ms)
✔ missing upstream usage remains unknown and does not become a free run or invented cost (43914.1637ms)
✔ missing raw check logs preserve unknown result and original sample denominator (37843.7607ms)
✔ source and model changes never silently replace an existing batch input (4016.6127ms)
✔ model revision change before preparing a fixed sample cannot silently use the replacement connection (3469.6471ms)
✔ HTTP evaluations require host token, original fingerprint and session ownership; worker API cannot bypass evaluation flow (15852.2952ms)
✔ sample queue cannot skip or reset attempted samples, and reports freeze further writes (28351.3525ms)
✔ stop cancels only current candidate and preserves partial results instead of replaying a request (19241.8626ms)
✔ restart recovers interrupted evaluations as unknown and keeps completed checks and plan identity (18515.5494ms)
✔ JSON migration survives process exit before commit and retains exact legacy backups (2448.6002ms)
✔ process exit after_state rolls back state, event and outbox together (2411.7073ms)
✔ process exit after_event rolls back state, event and outbox together (3563.6929ms)
✔ crash recovery at accepted retains execution identity and never replays side effects (4586.1776ms)
✔ crash recovery at queued retains execution identity and never replays side effects (2457.4333ms)
✔ crash recovery at approval_wait retains execution identity and never replays side effects (5309.6039ms)
✔ crash recovery at approved retains execution identity and never replays side effects (4652.8706ms)
✔ crash recovery at effect_unrecorded retains execution identity and never replays side effects (5075.5712ms)
✔ crash recovery at result_committed retains execution identity and never replays side effects (5916.6453ms)
✔ crash recovery at before_next_model retains execution identity and never replays side effects (2940.3553ms)
✔ state and event failures publish no success and preserve the last committed memory snapshot (752.356ms)
✔ tool result artifacts and event cursors retain project ownership across restart (1281.1919ms)
✔ token budgets pause before tools execute and continuation keeps original task and accumulated usage (1520.1666ms)
✔ a time budget interrupts an unresponsive model and persists a resumable run (1409.7539ms)
✔ a corrupted large result artifact cannot become a recovered success (1092.951ms)
✔ hard-link aliases cannot expose protected bytes through reads, proposals, inventory or automatic project rules (571.4529ms)
✔ ambiguous project hard links are explicit and ordinary independent file copies remain readable (423.5165ms)
✔ proposals preserve project bytes until review; apply backs up the exact original bytes (592.5944ms)
✔ new nested files are created only on apply and unchanged content creates no proposal (526.5125ms)
✔ stale reads and changed baselines are rejected without overwriting external edits (599.7996ms)
✔ file paths reject traversal, absolute paths, credentials and outside junctions (519.3554ms)
✔ read and search expose bounded text scopes and reject non UTF-8 / oversized files (563.2824ms)
✔ project-local junctions cannot expose or create protected Git and credential files (644.7768ms)
✔ Windows trailing-dot and trailing-space path aliases cannot bypass protected paths (565.8333ms)
✔ a starter is copied into user data once; restarting from another program version preserves existing projects and editable bytes (1429.1345ms)
✔ startup does not insert a demo or rewrite/reorder an existing user project registry (863.733ms)
✔ PNG bytes are fully decoded, metadata-free and bounded; JPEG EXIF orientation is applied to the sent PNG (106.9287ms)
✔ forged type, short headers, CRC damage, incomplete chunks, trailing payloads, animation and dimension bombs reject (30.2311ms)
✔ draft image ownership, withdrawal, reload and atomic Task binding retain original SHA and original bytes (1310.2654ms)
✔ a failed session transaction restores draft status and cannot leave a Run consuming the attachment (759.8287ms)
✔ draft limits, duplicate selection, source changes and cross-Task association reject before dispatch (1921.6173ms)
✔ demo and legacy paths refuse image dispatch without creating a Task or consuming the draft (1140.414ms)
✔ chat-completions sends one owned image per actual request, pairs original tools and verifies exact sealed bytes (3172.0126ms)
✔ responses sends one owned image per actual request, pairs original tools and verifies exact sealed bytes (2878.8227ms)
✔ mixed pool excludes unknown vision before lease and sends zero HTTP to the incompatible member (2618.4386ms)
✔ vision probe requires correct randomized pixel order and evidence expires when the connection revision changes (1038.8551ms)
✔ HTTP upload/read/withdraw uses session authorization, returns byte SHA and refuses unknown ownership (769.5628ms)
✔ image accounting excludes Base64 as text but seals every original request byte and reserves image input separately (3.1252ms)
✔ MCP registration stays disabled until an explicit real stdio probe and schema-based tool selection, with actual process cleanup (1924.3774ms)
✔ MCP probe holds the project operation lock until its owned process and report are settled (1521.988ms)
✔ MCP validates the declared JSON schema subset and rejects unsupported schemas before external dispatch (0.5898ms)
✔ actual model MCP call approves long schema-valid arguments, validates the sent catalog and stores one original result without leaking environment secrets (4252.0801ms)
✔ denying an MCP call never starts another process or executes the external tool (3652.0668ms)
✔ actual handshake drift rejects tools/call even if program entry hash is unchanged (4639.4303ms)
✔ cancelled dispatched external side effect stays unknown, closes its owned process and cannot replay on reopen (4157.9553ms)
✔ changed executable input, foreign project IDs and corrupted original contracts block enable and dispatch (1718.7996ms)
✔ registered MCP HTTP API uses fixed execution contracts and project-owned probe records with no model requests (5324.0488ms)
✔ unknown MCP pauses later tool calls and all project writers; fingerprinted acknowledgement clears only the review gates and explicit resume does not replay (7657.5313ms)
✔ MCP acknowledgement persistence failure restores both job and tool review gates; corrupt evidence cannot be acknowledged (4793.5786ms)
✔ actual host exit after an external effect restores unknown jobs without a fabricated report, allows source-bound review and never replays the call (3546.8749ms)
✔ changed enabled MCP programs can be explicitly disabled without launching or trusting their changed source (1538.8251ms)
✔ stdio preserves split UTF-8 frames and final replies drained after process exit (1489.4924ms)
✔ server-initiated model requests are refused instead of inheriting provider privileges (778.8996ms)
✔ invalid or over-budget protocol invalid closes only its owned process and is not retried (816.9193ms)
✔ invalid or over-budget protocol oversized closes only its owned process and is not retried (857.1342ms)
✔ cancel and timeout do not resend a dispatched request and pre-cancel has no request trace (851.7663ms)
✔ startup failure remains undispatched and has no process PID to terminate (466.5481ms)
✔ REA launch filters a fixture credential canary and rejects changed entry identity before a new connection (825.1782ms)
✔ schema1 connections remain readable without inventing observed capabilities or migrating original state (590.9699ms)
✔ protocol, unsupported capabilities and stream choice persist; client claims cannot forge verified evidence (678.1951ms)
✔ ability evidence belongs to exact connection and credential revision; unknown and unsupported reject required routing (0.2973ms)
✔ invalid declarations and inconsistent stream settings reject before registry, revision and credentials mutate (571.6857ms)
✔ model registry keeps original credentials/endpoint intact, rejects stale edits, and forgets only volatile keys on restart (835.1747ms)
✔ config import previews exact revisions, keeps secrets volatile, syncs only when enabled and preserves manual edits on conflicts (667.1254ms)
✔ real HTTP pool fails over 429 to another model and never repeats completed tools; attempts seal actual request bytes and Run identity (3118.9309ms)
✔ pool rotation and weights select members on new Runs; disabled and missing credentials are excluded without requests (3393.8152ms)
✔ 401 blocks the credential revision; updating the key restores availability without leaking error echoes (1208.0381ms)
✔ queued request cancellation and in-flight disable release only owned leases and do not dispatch another HTTP request (741.0114ms)
✔ compiled model-center HTTP API persists public profiles, blocks forged origins, and local operations never invoke a model (2175.013ms)
✔ model registry disk failures restore routing, pool profiles and credential references without partial updates (563.7484ms)
✔ model selection and saved configuration commit together and failures preserve the original mode and endpoint (574.6404ms)
✔ configuration imports roll back the entire batch and retain one usable preview after validation or commit failure (545.6834ms)
✔ HTTP quota retains native code/type/param/request id, bounded original and allowlisted headers, never repeats that connection (1513.1937ms)
✔ Retry-After above 90 seconds remains exact and a short local budget cannot trigger an early retry (1.2466ms)
✔ Retry-After date, expired date, huge value and invalid header retain explicit semantics (0.8606ms)
✔ unclassified 429 remains unspecified and every bounded actual retry retains its own original error (1319.0807ms)
✔ oversized HTTP error saves exactly the bounded prefix and labels its incomplete range (1670.1574ms)
✔ Chat SSE error preserves upstream fields, original failed event, partial tool and reported usage without executing or retrying (1321.168ms)
✔ Responses failed SSE retains original event/response identity and usage (1642.7267ms)
✔ Responses incomplete SSE retains original event/response identity and usage (2257.9978ms)
✔ unsupported successful Responses output retains original JSON and usage with a local protocol origin (2020.2479ms)
✔ HTTP 200 native error is still an upstream failure rather than a fabricated successful answer (0.6377ms)
✔ JSON-escaped credential echoes are redacted in the original preview and capture as well as parsed fields (0.4849ms)
✔ quota failover uses another independent pool member and never replays a failed stream (3362.0461ms)
✔ native parameters are isolated by pool member, preserve Responses state and bind the original output reservation (0.3293ms)
✔ unsupported native fields and invalid values reject before dispatch instead of silently dropping or overriding constraints (0.2872ms)
✔ native settings for legacy and registered profiles persist without changing credentials or old records and reject stale edits (1510.5156ms)
✔ compiled native settings API saves/restores exact parameters and rejects reserved fields without mutating the profile (2683.3494ms)
✔ legacy connection test sends its saved native fields and selected output limit field without changing the configuration (2682.2963ms)
✔ saved engine budgets verify all Chat/Responses route variants and reject a mismatched output allowance (1154.2366ms)
✔ owned CC Switch SQLite imports the selected TOML Responses wire_api and leaves unknown clients untouched (1084.3841ms)
✔ Responses executes one original tool through Runtime, carries encrypted Items, reports raw usage and seals exact wire bytes (3324.9812ms)
✔ mixed-protocol pool keeps both actual variants and changes route on 429 without re-executing completed tools (4477.6521ms)
▶ 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" (1752.4366ms)
✔ Reject {"body":{"id":"resp_owned","status":"failed","output":[{"type":"function_call","id":"fc_failed","cal (1772.9085ms)
✔ Reject {"body":{"id":"resp_owned","status":"completed","output":[{"type":"function_call","id":"fc_bad","cal (1656.7706ms)
✔ Reject {"body":{"id":"resp_owned","status":"completed","output":[{"type":"computer_call","id":"not-supporte (2165.9522ms)
✔ Responses incomplete, failed, missing call identity and unsupported hosted outputs never dispatch a tool (7348.8133ms)
▶ invalid JSON in a nonstream Chat or Responses tool group prevents every tool and pool replay
✔ chat-completions (2410.8756ms)
✔ responses (2300.1498ms)
✔ invalid JSON in a nonstream Chat or Responses tool group prevents every tool and pool replay (4711.4021ms)
✔ wire adapter preserves tool IDs but never passes Responses-private Items to Chat or another endpoint (0.1758ms)
✔ connection JSON import accepts explicit Responses protocol and continues to reject unknown protocols without reading client sessions (1582.2505ms)
✔ capability routing skips text-only members before acquiring a lease and records exact refusal in the original Run (2084.4675ms)
✔ all incompatible members reject with no network request, no lease, preserved Task and explicit non-dispatched ledger (1954.0352ms)
✔ explicit synthetic tool probe verifies only actual matching call, never executes it and loses verification on revision changes (1067.872ms)
✔ plain text probe cannot verify tools; wrong tool nonce and client supplied evidence cannot claim verified support (1141.1045ms)
✔ a disabled connection invalidates an in-flight probe even when toggle mutates its original registry object (961.0856ms)
✔ HTTP download preserves original malformed SSE bytes and artifact SHA instead of decoding invalid UTF8 (3277.2844ms)
✔ terminated SSE adapts empty optional tool arguments, retains the exact wire bytes and executes the validated batch once (4575.4135ms)
✔ reported SSE usage survives rejected required arguments "" without dispatching any tool (2804.9298ms)
✔ reported SSE usage survives rejected required arguments "{\"path\":" without dispatching any tool (2749.0545ms)
✔ reported SSE usage survives rejected required arguments "null" without dispatching any tool (2594.5568ms)
✔ a stream without DONE retains reported usage but cannot adapt or execute even an optional tool (1831.9862ms)
✔ a timeout after reasoning-only fragments creates one unknown receipt and never invents reported tokens (2214.895ms)
✔ real Chat SSE carries UTF8 and split tool JSON; Runtime executes only after DONE, stores partial progress and original usage (2823.5327ms)
✔ Responses typed SSE validates original item identity, terminal data, raw usage and tool result pair (2192.6118ms)
✔ incomplete network stream retains unknown attempt and never automatically switches pool or dispatches fragment tools (2640.9925ms)
✔ stopping an active HTTP stream closes only original request and keeps partial text without executing tools (3415.4473ms)
✔ explicit streaming probe uses actual SSE despite unknown declaration and binds its own original attempt (1795.7997ms)
✔ coalesced initial SSE events publish the latest partial state even when the upstream stays idle (2252.7303ms)
✔ stream timeout closes the original HTTP response, preserves redacted capture and never retries (1202.5994ms)
✔ stream idle timeout preserves incomplete tool arguments, records unknown use and never fails over (1367.9949ms)
✔ SSE accepts CR-only final event boundaries and retains exact invalid UTF8 bytes without dispatch (0.7736ms)
▶ stream parser rejects truncation, finish length, malformed JSON, duplicate sequence, changed final item and oversized data
✔ reject 0 (0.1699ms)
✔ reject 1 (0.1388ms)
✔ reject 2 (0.1309ms)
✔ reject 3 (4.9484ms)
✔ reject 4 (0.1272ms)
✔ reject 5 (0.1314ms)
✔ stream parser rejects truncation, finish length, malformed JSON, duplicate sequence, changed final item and oversized data (6.0848ms)
✔ older configurations receive bounded wait defaults without changing the original object (1.0016ms)
✔ active stream survives sixty seconds; byte activity resets idle but cannot extend total deadline (0.7847ms)
✔ silent stream has its own accurate deadline; caller cancellation and cleanup preserve their cause (1.2738ms)
✔ waiting for headers is bounded independently of stream-idle time (0.2255ms)
✔ wait settings persist, pin new runs and leave model pool, credentials and old run snapshots intact (854.6401ms)
✔ paged inspection finds a corrupted old artifact outside the recent fifty and retains original evidence without model/tool execution (5810.0026ms)
✔ inspection rejects ambiguous/out-of-range pagination and reports unknown tool/task identity without resolving it (1165.451ms)
✔ exact patches preserve BOM/CRLF, unicode and all unselected bytes and back up the original on apply (1140.106ms)
✔ several patches are matched against one original snapshot in any input order and keep LF (1559.8953ms)
✔ mixed line endings require exact anchors and preserve untouched formatting (831.9758ms)
✔ ambiguous, overlapping, missing and malformed patch anchors never create a proposal or write (615.871ms)
✔ stale original hashes and changes before apply preserve owner edits and mark conflicts (1153.3545ms)
✔ no-op BOM patches/full content produce no change and replacements cannot split Unicode bytes (1012.6194ms)
✔ an actual shell command keeps Node and UTF-8 output while omitting host-only credentials (1129.9979ms)
✔ real npm tests and build do not inherit arbitrary host credentials or Node injection settings (2198.4406ms)
✔ real Python tests and compilation keep their sealed inputs without host credentials (2849.9349ms)
✔ a process exiting between the Windows snapshot and owner query is absent only after a fresh query (1425.2934ms)
✔ owner observation errors remain errors while a process is present, replaced, or cannot be rechecked (1431.4389ms)
✔ cache controls require desktop authority, exact inputs and owned project identity while preserving task evidence (3135.6392ms)
✔ cache enablement and derived reuse survive a real host restart without becoming task recovery state (5557.5801ms)
✔ persistent parsing is project/SHA/config bound and produces identical semantic indexes after restart (1581.3534ms)
warning: in the working copy of 'README.md', LF will be replaced by CRLF the next time Git touches it
warning: in the working copy of 'main.ts', LF will be replaced by CRLF the next time Git touches it
warning: in the working copy of 'value.ts', LF will be replaced by CRLF the next time Git touches it
✔ same-size restored-mtime changes, additions, renames and deletions cannot reuse stale entries or import targets (1353.8922ms)
Switched to a new branch 'other'
✔ branch and commit changes invalidate even byte-identical sources (2273.7184ms)
✔ corrupt, malformed and schema-invalid records fall back to fresh parsing (1117.4445ms)
✔ disable and clear persist without touching sources, sessions, configuration or original artifacts (1046.5964ms)
✔ cache capacity bounds disk and memory by bytes and LRU project count (932.1418ms)
✔ incomplete scans do not load or save cached entries; unavailable files leave the current result scope (918.3778ms)
✔ project and cache junctions cannot expose or delete an outside directory (1025.3755ms)
✔ concurrent file changes refuse a snapshot and clear during pending work cannot repopulate it (988.9604ms)
✔ file access mode changes invalidate parsed material even when bytes are unchanged (1038.1373ms)
✔ tracked directory replaced by a junction invalidates access scope and never uses its old cache (1754.5444ms)
✔ JS/TS/React AST indexes actual declarations, exports, imports and verified source lines (667.9837ms)
✔ default expressions, named default exports and declared functions retain their AST evidence (549.3554ms)
✔ incremental refresh invalidates changed/deleted symbols and recalculates reused import targets (1133.0476ms)
✔ Python candidates exclude comments and multiline strings and declare limited evidence (767.2294ms)
✔ indexes isolate project identity and root and evict bounded caches without sharing results (898.4854ms)
✔ Git ignore, credential paths and links stay outside the index and no project files are executed (1496.9277ms)
✔ malformed, binary, non-UTF8 and bounded files expose partial coverage while good files remain searchable (1278.5963ms)
✔ index query rejects invalid limits/types and does not imply reference or alias resolution (984.8553ms)
✔ read-only engineering inspection disables configured Git clean filters and preserves index bytes (2268.8639ms)
✔ Python multiline declarations retain candidate scope without interpreting expressions as declarations (700.7446ms)
✔ TypeScript module candidates support import-equals and MTS/CTS while opaque imports stay explicit (1269.1041ms)
✔ actual HTTP requests reuse only this project source-checked memory as redacted data and preserve sent evidence after archive/restart (2656.3872ms)
✔ actual HTTP request excludes archived, timed-out and foreign-project memories (3719.7233ms)
✔ actual next HTTP request removes expired memory after a tool observes a source change without replaying that tool (4318.8344ms)
✔ incomplete inventory and excluded-file SHA changes cannot enter actual HTTP memory context or inspect_session snapshots (2094.4823ms)
✔ corrupt original artifact expires its memory and cannot reach an HTTP model request (2024.4863ms)
✔ memory JSON is bounded in escaped UTF-8 bytes and entry count, skips oversized entries and retains complete multilingual values (3665.8303ms)
✔ optional memory yields to the complete goal and original code evidence under the whole-request budget (1998.6483ms)
✔ changed sealed memory artifacts or binding fields fail verification while the original request remains immutable (1844.0539ms)
✔ a malformed project declaration omits optional memory while the repair Task still reaches the owned HTTP service (1200.1329ms)
✔ an expiration persistence failure prevents dispatch and is not hidden as an optional scan failure (883.6247ms)
✔ project memories persist across restart, keep source provenance, and are visible across tasks (1635.9745ms)
✔ source changes expire current project memories and keep the original evidence (1266.1889ms)
✔ incomplete source scans mark current memories unverified without expiring or reusing them (1422.9025ms)
✔ explicit file provenance is rechecked even when the inventory excludes that path (1437.2052ms)
✔ execution snapshots do not expose project memories before their source scope is validated (2193.8293ms)
✔ project memory records stay isolated and archive requires the owning session (1074.751ms)
✔ project memory HTTP API exposes read-only source status and explicit writes (2441.6109ms)
✔ two Purge source versions lock nine original inputs without executing JavaScript or postinstall and remain disabled by default (844.2226ms)
✔ incomplete, unsupported, relative and oversized sources cannot enable the experiment (657.7747ms)
✔ invalid source fingerprints and invented privilege or retry fields fail before any parent Task is created (2570.4351ms)
✔ source edits between detection and plan are rejected without creating a parent Task (3222.3465ms)
✔ source edits after sealing cannot launch a child and do not reset its cumulative budget (8276.4189ms)
✔ literal experiment text is sealed into the parent, worker and each new Run while actual host approvals and saved results remain required (25754.7297ms)
✔ source changes after a model response stop before any native tool is accepted, retaining the actual usage once (10461.2123ms)
✔ source changes while approval is pending invalidate the old permission before an engineering check starts (13391.4225ms)
✔ altered text, interpolation, section order and policy flags cannot reuse a sealed experiment (2007.9365ms)
✔ the observed SDK system-message layout and the explicit system field both require exactly one effective section, never a user or tool copy (1919.9438ms)
✔ a missing or duplicated experimental section in the actual input is rejected before any model dispatch (18460.1504ms)
✔ experimental prompt overhead still participates in the original complete-input budget guard (10399.9851ms)
✔ source detection HTTP remains local-authenticated and read-only with no configured model requests (2076.9697ms)
✔ Python HTTP plans keep npm defaults, require owned approval and preserve actual results across restart (9905.3089ms)
✔ an isolated three-module Python Task repairs, checks and delivers a source-bound Git report without extra model calls (34465.0931ms)
✔ Python static inspection does not import project code or probe a project interpreter before approval (1170.7919ms)
✔ a real three-module Python failure, reviewed repairs, tests and compilation retain exact structured evidence (4721.523ms)
✔ unittest counts repeated execution ordinals and subtest skips rather than deduplicating display names (2887.4118ms)
✔ unittest empty, skipped, expected failure and class initialization errors never verify a task (7711.4359ms)
✔ stdout claims cannot replace actual unittest results and an early exit cannot forge a successful collector (2411.0109ms)
✔ compilation reads actual Python source bytes without executing top-level project code (3318.3359ms)
✔ Node and Python mixed projects retain the existing npm default and require an explicit Python selection (2157.7698ms)
✔ changed project interpreters and forged Python plan bindings execute no check (1437.249ms)
✔ malformed TOML and invalid Python test paths are reported without expanding npm check scope (1086.1843ms)
✔ stopping a real Python check terminates its child and keeps the completed result as cancelled (1666.5808ms)
✖ real pytest pass, failure, skip, collection errors and teardown errors use collected case outcomes (1615.6318ms)
✔ a real project virtual environment is detected statically and its actual version is recorded only during the approved check (7014.4484ms)
✖ a missing interpreter and repeated pytest item objects cannot create a false verification pass (2424.1115ms)
✔ continuation reads advance through historical pages without a false pause while repeated current-Run pages still pause (2.113ms)
✔ a new tool page replaces old latest-feedback space and still obeys its escaped protocol budget (0.6037ms)
✔ budgeted artifact pages count escaped protocol bytes and reconstruct original multilingual JSON without losing characters (9.0118ms)
✔ known artifact coverage detects semantic repeats despite changing limits, and advancing pages or direct evidence restore progress (0.3367ms)
✔ actual large artifact reads are present in the next bounded HTTP request and next_offset progresses without redoing tools (3934.6312ms)
✔ a real repeated artifact loop pauses the same Task after recorded results, survives reload and only resumes explicitly (6658.4857ms)
✔ actual HTTP continuation after reload sends distinct historically covered pages without false no_progress or replay (5746.0237ms)
✔ an installed app outside the source drive resolves registered SDK and Purge through actual input verification (865.0808ms)
✔ standalone registration resolves the upstream metadata directory and missing upstream falls back to the other registered source (747.9038ms)
✔ registered hints keep explicit roots, environment and existing app-relative candidates in their original priority (1192.9511ms)
✔ unknown IDs, duplicate registered IDs, wrong versions and relative or inconsistent metadata are not lookup candidates (800.6296ms)
✔ oversized, malformed, invalid UTF-8 and non-file registry inputs cannot silently enable a source (671.5816ms)
✔ a linked registry directory is rejected without reading or deleting its external source (747.7255ms)
✔ matching research declarations do not replace actual installed version and critical-source checks (885.69ms)
✔ context APIs retain exact request identity, paginate archived groups and reject foreign sessions/runs (3914.4029ms)
✔ delivery reports verify request body, sealed plan and original archived bytes, and show corruption as needing verification (2680.513ms)
✔ direct code evidence remains selected while recent historical-index browsing is archived within the complete request budget (2728.6385ms)
✔ one long user turn becomes bounded complete groups with exact UTF-8 history and recent JSON arguments (5475.4329ms)
✔ tool definitions, exact system bytes, output reserve and remaining Run/Task budget are accounted before requests (1275.3617ms)
✔ oversized required user instructions pause intact and can be explicitly resumed after increasing the window (2220.3285ms)
✔ all task instructions, constraints and acceptance remain in the runtime system instead of an eight-item or byte prefix (1490.5145ms)
✔ a newest manageable multi-tool group stays exact even beyond the working target; oversized earlier groups are archived (1321.1491ms)
✔ invalid orphan result is rejected before archive or network preparation (915.6754ms)
✔ invalid missing result is rejected before archive or network preparation (1146.3172ms)
✔ invalid wrong result is rejected before archive or network preparation (1033.4983ms)
✔ invalid duplicate call is rejected before archive or network preparation (822.6072ms)
✔ a group larger than the artifact limit uses verified Unicode-safe chunks without replacing tool JSON (2399.5158ms)
✔ changed or missing archived evidence stops reuse and is never silently recreated (1236.8596ms)
✔ transaction failure prevents publishing a plan and any corresponding model request (1086.5394ms)
✔ the Provider rejects body mutation after a sealed plan before calling HTTP (866.0377ms)
✔ HTTP failure and explicit same-Task resume reuse exact completed results and request hashes without repeating a command (10400.3701ms)
✔ a real process exit after plan persistence marks interruption and never issues an automatic model request (1821.5907ms)
✔ deep Windows data directories keep REA output within the external path allowance and retain owned evidence (3396.4738ms)
✔ user research registration creates an owned Task without model requests and seals the exact artifact identity (1169.1415ms)
✔ complete research binds original Evidence and Bundle, pages exact fields and reuses verified results after reload (2515.1512ms)
✔ artifact changes before analysis refuse dispatch and changes during analysis preserve unverified original responses (3417.7216ms)
✔ owned cancellation closes the connection, concurrent and automatic retries are refused, explicit new analysis preserves history (2423.2944ms)
✔ a dispatched transport timeout stays unknown and reload never launches or kills a saved PID (2289.5798ms)
✔ corrupt Evidence parts and changed binding records are refused instead of silently re-analysed (2908.2931ms)
✔ research documents over the single-artifact limit preserve exact Unicode bytes with owned verifiable parts (1306.3644ms)
✔ research tools are only sent for registered Task artifacts and actual HTTP uses the sealed tool contract without replay (7077.4684ms)
✔ file identity rejects directory and invalid formats without running a target (809.049ms)
✔ a stop from the event loop interrupts immediate model replies before later tools (2176.8346ms)
✔ a stop after the first tool preserves a complete interrupted multi-call protocol group (2335.6646ms)
✔ the compatible HTTP loop preserves call IDs, validates errors and waits before executing commands (6878.097ms)
✔ denial prevents command execution and produces a paired denied result (2615.7049ms)
✔ stopping during approval interrupts pending calls without executing or replaying them (1799.7487ms)
✔ stopping a running fixture command terminates its own process tree and retains uncertain state (2408.8892ms)
✔ command output decodes split UTF-8 and excludes the workbench key from child environment (1175.5997ms)
✔ model-list testing uses a tentative key without changing the saved connection and redacts service errors (7.6306ms)
✔ truncated or repeated model calls fail before execution; tool input validation rejects unknown capabilities (3634.3663ms)
✔ request projection keeps exact tool pairs and complete recent user instructions while preserving disk history (1535.7109ms)
✔ local skill revisions require explicit enable, retain original text, survive restart and stay within the selected project (1291.5208ms)
✔ registration rejects non-SKILL paths, oversized/non-UTF8 files, duplicate sources and wrong field shapes without running project code (761.2608ms)
✔ actual HTTP models receive only enabled current-project metadata, then a redacted owned skill page with complete tool pairing (3405.9725ms)
✔ source edits after a catalog request prevent the read and appear as changed on the next request without replay (3253.569ms)
✔ disabled and foreign skills cannot be read by guessing their IDs from another project (2535.1776ms)
✔ skill catalog takes remaining escaped JSON budget and cannot evict the user goal or current code (1355.265ms)
✔ multilingual skill body pages fit actual tool JSON bytes and every page uses the same immutable artifact (4492.7517ms)
✔ registry failures roll back enable and refresh state; corrupted snapshots block requests before network (1668.8444ms)
✔ saved skill catalogs retain SHA-bound historical evidence and reject corrupt catalog blobs (2292.5203ms)
✔ HTTP skill administration enforces project scope, revision checks and idle mutation without model calls (3146.5742ms)
✔ JSON state keeps an atomic backup and preserves corrupt primary files (1104.2701ms)
✔ tool-pair recovery fills missing results and drops orphan or duplicate results (0.4889ms)
✔ restart pauses unfinished work, retains proposals and never replays a command (1075.462ms)
✔ public snapshots carry increasing revisions without the private provider transcript or config key (1070.194ms)
✔ a checkpoint seals all task requirements and original results without a model request and explicit continuation keeps the same Task without replay (3923.4834ms)
✔ checkpoint sources, requirements, plan and original tool changes block continuation and leave original checkpoint bytes intact (3780.2577ms)
✔ unknown tool effects and pending proposals stay blocked in recovery packages rather than being changed into completed work (3616.1431ms)
✔ checkpoint and referenced result corruption are rejected before dispatch and historical blobs are never rebuilt (3233.9604ms)
✔ checkpoint ownership, exact confirmation and concurrent operations reject foreign or stale resumes (3438.7118ms)
✔ Git identity changes and altered plan proofs cannot be accepted as the sealed recovery state (4245.7038ms)
✔ checkpoint records survive SQLite reopen and do not trigger automatic model or tool replay (4282.3045ms)
✔ incomplete scans and in-progress project activity refuse checkpoint creation while private credentials are redacted from saved requirements (4342.918ms)
✔ oversized recovery evidence is refused before reading or publishing the checkpoint (3217.6697ms)
✔ compiled checkpoint HTTP APIs bind current Task/Run, isolate sessions and only resume on an explicit matching confirmation (7234.721ms)
✔ actual process exit before checkpoint commit retains only durable records without replay (2150.8735ms)
✔ actual process exit after checkpoint commit retains only durable records without replay (3195.8588ms)
✔ operational questions are diagnostic while explicit repair requests remain execution instructions (1.0159ms)
✔ file delivery requires a positive instruction, not saved evidence or a prohibited write (1.812ms)
✔ an unsubstantiated text claim cannot complete a requested file delivery (1590.7965ms)
✔ HTML without tests ends at reviewed file delivery and never offers automatic rechecking (2641.5891ms)
✔ diagnostic answers retain the unfinished task, limits and check result, and enforce read-only tools (4166.6781ms)
✔ review continuation is a system action and adds no testing demand or fake user instruction (2720.0356ms)
✔ a provider cannot execute tests against an explicit no-test task even if it ignores advertised tools (2329.6254ms)
✔ a budget pause becomes visible only after its input snapshot can be sealed with the original Run (1823.0416ms)
✔ alternating duplicate file reads and listing files still pauses without treating new pages as duplicates (20119.9296ms)
✔ task-plan HTTP mutations bind current Task/Run and retain original evidence through real host restart (4895.3303ms)
✔ concurrent stale HTTP plan revisions cannot replace conditions or attach proofs to a different plan (3330.7974ms)
✔ a real escaped long-file tool message keeps its proof identity within a tight request budget (1171.0718ms)
✔ task plans reject fabricated completion, invalid scopes and cyclic or missing dependencies (0.4751ms)
✔ sealed plans retain goal, contract hash and original steps through SQLite reopen without executing tools (1169.4203ms)
✔ proofs require original owned reads, honor dependencies and reject altered invocation arguments and bytes (1538.364ms)
✔ new user requirements invalidate current proofs and reseal the same original operations in a new revision (1757.1637ms)
✔ applied change-set evidence binds reviewed current files and invalidates dependent steps after outside edits (2344.6347ms)
✔ actual npm checks are verified by original counted logs and current source, with failure history retained (6895.9738ms)
✔ real desktop and mobile browser proofs bind required actions and PNGs, rejecting source changes, a later failed plan and changed builds (12422.8183ms)
✔ Git delivery blocks incomplete plans and changed preview bindings, preserving HEAD and owner index before a fresh verified commit (15113.4068ms)
✔ the runtime supplies owned read IDs and completes a plan only after original step evidence verification (1915.1764ms)
✔ premature model completion pauses the original Task and resume preserves the existing plan without replay (2036.1087ms)
✔ plan operations participate in project activity and orderly shutdown (574.3688ms)
✔ actual process exit before plan save restores only committed plan evidence without replay (1254.3639ms)
✔ actual process exit after plan save restores only committed plan evidence without replay (1406.0813ms)
✔ actual process exit before proof save restores only committed plan evidence without replay (1670.9716ms)
✔ actual process exit after proof save restores only committed plan evidence without replay (1718.2701ms)
✔ the latest oversized session result reaches the next request together with already read source modules (2791.4143ms)
✔ every call in a large latest group keeps its actual outcome, including denied and unknown results (2004.6174ms)
✔ a small complete latest tool group is sent once and needs no duplicate feedback view (1309.3041ms)
✔ real HTTP repeated state inspection pauses after three identical engineering outcomes and preserves tool pairing (3811.2572ms)
✔ a live SQLite host with the actual 1.9 artifact object upgrades, keeps its original blobs and verifies real browser PNG bytes (3901.9245ms)
✔ failed upgrade restores the original artifact object, provider and request handler (1591.4161ms)
✔ claiming the same runtime owner twice preserves its original metadata and every database row (664.6524ms)
✔ sessions without controlled-engine history retain their prior export shape and never expose another session record (617.5151ms)
✔ live workers, controlled drivers, or MCP operations refuse an upgrade before changing handlers, managers or database rows (751.1707ms)
✔ an idle upgrade loads the newly introduced engine record map, resets obsolete managers and preserves original rows, sessions and provider objects (664.9408ms)
✔ an installation failure restores old managers, original record maps, handlers and all existing database rows (1424.1686ms)
✔ the owned controlled driver remains an active session and project even after outer task activity has ended (655.7593ms)
✔ release and upgrade help or invalid arguments have no file or network setup side effects in an independent directory (810.1058ms)
✔ hot upgrade adopts a retained virtual data root without reopening SQLite or changing rows, Provider or routing (1150.846ms)
✔ failed adoption restores the original virtual data-root name and original open database (526.7228ms)
✔ hot upgrade refreshes retained model and importer methods while preserving credential and routing objects (547.1134ms)
✔ hot upgrade refuses active probes and restores retained model methods if installation fails (582.5602ms)
✔ a directly started compiled host upgrades twice with the same PID, sessions, applied bytes and volatile provider (3509.3523ms)
✖ the running 0.1.0 process upgrades in place and retains its project, session, files and volatile model connection (2537.371ms)
✔ HTTP worker identities require a current sealed plan, reject overrides and close unused plans without touching source (17229.5015ms)
✔ HTTP unknown recovery only accepts the original evidence SHA and preserves changed source for explicit retention (20540.3796ms)
✔ verified isolated outputs form a new parent batch, rerun real checks and commit only reviewed source (53897.324ms)
✔ overlapping worker paths are explicit conflicts even if each independent task checks successfully (61428.3771ms)
✔ source edits after isolated work preserve the owner file and invalidate collection without proposing a merge (55987.3569ms)
✔ changed child bytes after parent preview block the entire batch before any parent write (59973.5539ms)
✔ corrupt original check logs cannot become collected outputs or a parent review (32140.9885ms)
✔ stopping a parent accounts late responses and executes none of their proposed tools (16099.3047ms)
✔ unreported interrupted requests reserve estimated usage and the total deadline cannot reset on resume (16259.6429ms)
✔ a deadline before worker request dispatch records no invented usage and cannot reset on resume (19258.9566ms)
✔ a response exceeding the cumulative token allocation stops before registering or executing its tools (26412.2945ms)
✔ cumulative tool allocation rejects the entire returned batch before any tool side effect (28485.9812ms)
✔ single response 16 tool ceiling rejects the entire returned batch before any tool side effect (27727.2244ms)
✔ consumed request steps remain consumed across explicit worker resumes (27976.5232ms)
✔ a saved completed response with missing usage is conservatively accounted exactly once during recovery (17741.063ms)
✔ an unexecuted second workspace plan resumes its original identity without recreating the first checkout (21759.7041ms)
✔ a verified interrupted parent proposal resumes sealing without duplicating or replacing its original change (50599.2994ms)
✔ recovery attaches a sealed parent batch saved before its group reference without creating another batch (44962.3408ms)
✔ explicitly retaining an unexecuted parent merge cancels its batch and frees the next parent goal (43229.6316ms)
✔ unknown collaboration blocks ordinary proposals but permits only original batch recovery and rollback (48174.5307ms)
✔ after a restart the batch application lazily reinstates the worker result guard before any parent write (34346.7761ms)
✔ unknown original tool inputs must match before acknowledgement, and reviewed checkpoint IDs preserve the unknown fact (14746.4223ms)
✔ a parent stop observes and waits for a manually approved child check and its real descendant process (14171.6275ms)
✔ a real host crash preserves unknown checks, refuses a still living owned PID and never replays a model or tool (24336.5694ms)
✔ a persisted child initialization resumes the original control run after a failure before its worker record (12659.6292ms)
✔ failed request intent persistence never dispatches a worker request or changes source (13338.3247ms)
✔ failed usage persistence rolls back the uncommitted usage and conservatively reserves one unknown request without tools (13541.9483ms)
✔ a child file change after observation prevents acknowledgement even when the tool inputs did not change (12935.9206ms)
✔ worker scopes and allocated budgets reject aliases, extra permissions and aggregate overflow (680.1364ms)
✔ two sequential Git workspaces freeze a new parent goal, separate owned tasks and preserve all source bytes (14408.0475ms)
✔ real isolated changes, approvals and npm checks produce original owned outputs, not model completion claims (34015.5506ms)
✔ forged worker tools cannot execute terminal or Git even when absent from the advertised tool list (14732.4927ms)
✔ a completion-only model pauses each worker without fabricated changes or checks (17204.1421ms)
✔ restart retains frozen scope and marks interrupted parent and workers unknown without sending requests (11853.834ms)
✔ small code and actual failed command stay visible when their original multi-tool group cannot fit (2177.8755ms)
✔ mixed group retains six distinct small file observations with exact multilingual text and bounded request bytes (5967.8503ms)
✔ working observations are bound to the current Task and stop treating a pre-application file revision as current evidence (1462.4365ms)
✔ a sealed observation rejects changed or missing original result artifacts instead of recreating evidence (1131.9956ms)
✔ a real passed-check observation is excluded after a source edit invalidates the check, while its historical sent evidence remains valid (4882.075ms)
✔ budgeted file pages reconstruct UTF-8 text including long single lines and surrogate boundaries (4.1619ms)
✔ actual HTTP multi-tool reads survive an oversized runtime snapshot and the sealed evidence survives reopening (12891.0637ms)
✔ actual HTTP file pagination advances over a long file with each returned page visible in the next request (2382.9151ms)
✔ later mixed history pages cannot displace the actual module sources and complete requirements (6528.9834ms)
✔ a later narrow page of the same file revision cannot evict the wider actual observation (3205.633ms)
✔ historical read-artifact navigation points to the original artifact and next page, never its result wrapper (839.4406ms)
✔ a 1.6.1 full-provenance sent view remains verifiable after introducing the compact view (1144.814ms)
✔ actual HTTP requests keep four observed modules after a later large history read and preserve all source bytes (10596.2157ms)
✔ workspace HTTP ownership, reviewed creation and reopen survive restart without changing source files (9573.5219ms)
✔ isolated model Task applies a batch, executes real checks and commits at the Git review boundary with a source-bound report (26719.7595ms)
warning: in the working copy of 'src/空 格.mjs', LF will be replaced by CRLF the next time Git touches it
✔ worktree checkout disables hooks, fsmonitor and all filter command variants while honoring built-in EOL conversion (6056.8042ms)
✔ an ignored external file at checkout is observed and never accepted as pristine creation (4711.2119ms)
✔ outside content hidden by assume-unchanged cannot pass workspace verification and blocks further source execution (4262.4009ms)
✔ workspace preparation refuses known shared-repository activity and preserves archived child sessions (4994.8604ms)
✔ compact checkout identities create in deep Unicode source directories while retaining the human branch name (5047.9807ms)
✔ Windows Git metadata path limits reject creation before publishing a branch or changing source bytes (2364.9339ms)
✔ a pending source snapshot participates in shutdown and completes before storage can close (2107.1135ms)
✔ source configuration values stay private and a configuration change invalidates a saved review (2547.2737ms)
✔ workspace review rejects excluded dependency before any checkout (2205.538ms)
✔ workspace review rejects large committed blob before any checkout (2338.8954ms)
✔ workspace review rejects case collision before any checkout (2647.5032ms)
✔ workspace review rejects committed symbolic link before any checkout (2279.7382ms)
✔ workspace review rejects external config include before any checkout (1591.8258ms)
✔ actual process exit at workspace intent preserves source and reconciles without creation replay (4168.4634ms)
✔ actual process exit at workspace git_started preserves source and reconciles without creation replay (6324.1667ms)
✔ actual process exit at workspace git_finished preserves source and reconciles without creation replay (5785.0331ms)
✔ actual process exit at workspace verifying preserves source and reconciles without creation replay (6486.7424ms)
✔ actual process exit at workspace registering preserves source and reconciles without creation replay (6387.857ms)
✔ actual process exit at workspace project_registered preserves source and reconciles without creation replay (7855.8549ms)
✔ actual process exit at workspace session_registered preserves source and reconciles without creation replay (7284.255ms)
✔ actual process exit at workspace ready preserves source and reconciles without creation replay (6312.3676ms)
✔ outside changes at project registration remain unknown and preserve the already registered child identity (6250.8201ms)
✔ managed worktree preserves source index and dirty files, opens a distinct project and binds Run identity (4737.6233ms)
✔ stale review, branch/path collisions and forged ownership do not run worktree creation (2940.806ms)
✔ workspace project verification is read-only and rejects physical Git identity changes (4225.4814ms)
ℹ tests 649
ℹ suites 0
ℹ pass 646
ℹ fail 3
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 1616111.3541
✖ failing tests:
test at tests\python-checks.test.mjs:171:1
✖ real pytest pass, failure, skip, collection errors and teardown errors use collected case outcomes (1615.6318ms)
AssertionError [ERR_ASSERTION]: ModuleNotFoundError: No module named 'pytest'
'failed' !== 'passed'
at TestContext.<anonymous> (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/tests/python-checks.test.mjs:183:12)
at process.processTicksAndRejections (node:internal/process/task_queues:104:5)
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: 'passed',
operator: 'strictEqual',
diff: 'simple'
}
test at tests\python-checks.test.mjs:202:1
✖ a missing interpreter and repeated pytest item objects cannot create a false verification pass (2424.1115ms)
AssertionError [ERR_ASSERTION]: ModuleNotFoundError: No module named 'pytest'
'failed' !== 'passed'
at TestContext.<anonymous> (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/tests/python-checks.test.mjs:209:10)
at process.processTicksAndRejections (node:internal/process/task_queues:104:5)
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: 'passed',
operator: 'strictEqual',
diff: 'simple'
}
test at tests\upgrade.test.mjs:38:1
✖ the running 0.1.0 process upgrades in place and retains its project, session, files and volatile model connection (2537.371ms)
Error: The local inspector port belongs to another process.
at upgradeRunning (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/scripts/upgrade-running.mjs:74:87)
at process.processTicksAndRejections (node:internal/process/task_queues:104:5)
at async TestContext.<anonymous> (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/tests/upgrade.test.mjs:70:18)
at async Test.run (node:internal/test_runner/test:1409:7)
at async Test.processPendingSubtests (node:internal/test_runner/test:974:7)
@@ -0,0 +1,45 @@
✔ HTTP quota retains native code/type/param/request id, bounded original and allowlisted headers, never repeats that connection (1491.9422ms)
✔ Retry-After above 90 seconds remains exact and a short local budget cannot trigger an early retry (1.3138ms)
✔ Retry-After date, expired date, huge value and invalid header retain explicit semantics (0.9495ms)
✔ unclassified 429 remains unspecified and every bounded actual retry retains its own original error (1109.1278ms)
✔ oversized HTTP error saves exactly the bounded prefix and labels its incomplete range (1335.1775ms)
✔ Chat SSE error preserves upstream fields, original failed event, partial tool and reported usage without executing or retrying (1346.671ms)
✔ Responses failed SSE retains original event/response identity and usage (1275.4702ms)
✔ Responses incomplete SSE retains original event/response identity and usage (1313.8146ms)
✖ unsupported successful Responses output retains original JSON and usage with a local protocol origin (1202.1233ms)
✔ HTTP 200 native error is still an upstream failure rather than a fabricated successful answer (0.9673ms)
✔ quota failover uses another independent pool member and never replays a failed stream (1664.322ms)
✔ native parameters are isolated by pool member, preserve Responses state and bind the original output reservation (1.2901ms)
✔ unsupported native fields and invalid values reject before dispatch instead of silently dropping or overriding constraints (1.4361ms)
✔ native settings for legacy and registered profiles persist without changing credentials or old records and reject stale edits (928.0481ms)
✔ compiled native settings API saves/restores exact parameters and rejects reserved fields without mutating the profile (2188.2182ms)
ℹ tests 15
ℹ suites 0
ℹ pass 14
ℹ fail 1
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 13988.257
✖ failing tests:
test at tests\model-fidelity.test.mjs:134:1
✖ unsupported successful Responses output retains original JSON and usage with a local protocol origin (1202.1233ms)
AssertionError [ERR_ASSERTION]: Expected values to be strictly equal:
+ actual - expected
+ 'upstream'
- 'local'
at TestContext.<anonymous> (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/tests/model-fidelity.test.mjs:138:10)
at process.processTicksAndRejections (node:internal/process/task_queues:104:5)
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: 'upstream',
expected: 'local',
operator: 'strictEqual',
diff: 'simple'
}
@@ -0,0 +1,18 @@
✔ full candidate validation returns the exact loaded-input host identity and rejects undeclared or altered files (3452.6201ms)
✔ installation requires separate absolute program/data directories and a candidate in its own versions namespace (3798.3109ms)
✔ staging retains the original data and pointer without querying or restarting a host (4847.9939ms)
✔ a post-pointer failure restores every original pointer byte and persistent row; a first-install failure removes only its published pointer (8746.0206ms)
✔ changed active manifest and concurrent pointer changes are rejected without overwriting the unknown record (4701.9443ms)
✔ an unregistered non-HTTP listener blocks preflight even when service-info is absent (2363.6739ms)
✔ uninstall archives its original pointer and preserves every data row/artifact instead of recursively deleting data (6695.616ms)
✔ preflight rejects an unrelated program root or changed native bootstrap while preserving its original bytes (1462.7646ms)
✔ a changed owned bootstrap follows its candidate and a post-pointer fault restores both original bytes (9878.0743ms)
✔ an external bootstrap change after preflight is preserved and never overwritten as an update or rollback (8600.3132ms)
ℹ tests 10
ℹ suites 0
ℹ pass 10
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 54644.1295
@@ -0,0 +1,25 @@
✔ HTTP quota retains native code/type/param/request id, bounded original and allowlisted headers, never repeats that connection (3430.5388ms)
✔ Retry-After above 90 seconds remains exact and a short local budget cannot trigger an early retry (1.2555ms)
✔ Retry-After date, expired date, huge value and invalid header retain explicit semantics (0.9076ms)
✔ unclassified 429 remains unspecified and every bounded actual retry retains its own original error (1912.3747ms)
✔ oversized HTTP error saves exactly the bounded prefix and labels its incomplete range (1566.9352ms)
✔ Chat SSE error preserves upstream fields, original failed event, partial tool and reported usage without executing or retrying (1719.405ms)
✔ Responses failed SSE retains original event/response identity and usage (2968.8225ms)
✔ Responses incomplete SSE retains original event/response identity and usage (2425.7177ms)
✔ unsupported successful Responses output retains original JSON and usage with a local protocol origin (1698.345ms)
✔ HTTP 200 native error is still an upstream failure rather than a fabricated successful answer (0.598ms)
✔ JSON-escaped credential echoes are redacted in the original preview and capture as well as parsed fields (0.445ms)
✔ quota failover uses another independent pool member and never replays a failed stream (2999.8504ms)
✔ native parameters are isolated by pool member, preserve Responses state and bind the original output reservation (0.331ms)
✔ unsupported native fields and invalid values reject before dispatch instead of silently dropping or overriding constraints (0.303ms)
✔ native settings for legacy and registered profiles persist without changing credentials or old records and reject stale edits (1644.0071ms)
✔ compiled native settings API saves/restores exact parameters and rejects reserved fields without mutating the profile (2675.1196ms)
✔ legacy connection test sends its saved native fields and selected output limit field without changing the configuration (2300.5915ms)
ℹ tests 17
ℹ suites 0
ℹ pass 17
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 25549.268
@@ -0,0 +1 @@
{"success":true,"retained":{"originalRows":"exact","originalExports":"exact","originalArtifactBytes":"exact","originalModelCenter":"exact"}}
@@ -0,0 +1,70 @@
{
"portableRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-win-x64-20261101141644779",
"sourceRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-source-20261101141644779",
"portable": {
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-win-x64-20261101141644779.zip",
"bytes": 629123292,
"sha256": "de760fe8fb0f4dc354e34365d813cb1866efb34b110769c9980b9c744f36d93a"
},
"source": {
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-source-20261101141644779.zip",
"bytes": 629305819,
"sha256": "a43dad2a7e638046d237b0142b07725bf1b57687b8d0823ef63f3367d8a3da22"
},
"installer": {
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-win-x64-20261101141644779-Setup.exe",
"bytes": 588262745,
"sha256": "c30c362c58641749b884d47c4ed906063074024eab19e8e494cd81afa91c0141",
"compiler": {
"version": "6.3.1",
"sha256": "3b22d36fc7e7c421d8caf780008ecdb42b14b3f0371b9bafe3cad64e6648d9cf"
},
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\installer-build-837235a2-b71b-4e64-8e60-f97c903969d9\\installer.iss",
"version": "1.20.3",
"candidateDirectory": "Agent-Workbench-1.20.3-win-x64-20261101141644779",
"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": 19638,
"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 +1,21 @@
✔ Python static inspection does not import project code or probe a project interpreter before approval (872.0865ms)
✔ a real three-module Python failure, reviewed repairs, tests and compilation retain exact structured evidence (3773.781ms)
✔ unittest counts repeated execution ordinals and subtest skips rather than deduplicating display names (2399.8299ms)
✔ unittest empty, skipped, expected failure and class initialization errors never verify a task (5881.8753ms)
✔ stdout claims cannot replace actual unittest results and an early exit cannot forge a successful collector (2141.1998ms)
✔ compilation reads actual Python source bytes without executing top-level project code (2145.4775ms)
✔ Node and Python mixed projects retain the existing npm default and require an explicit Python selection (2042.5217ms)
✔ changed project interpreters and forged Python plan bindings execute no check (1197.6955ms)
✔ malformed TOML and invalid Python test paths are reported without expanding npm check scope (1018.8024ms)
✔ stopping a real Python check terminates its child and keeps the completed result as cancelled (1480.3634ms)
✔ real pytest pass, failure, skip, collection errors and teardown errors use collected case outcomes (11854.8794ms)
✔ a real project virtual environment is detected statically and its actual version is recorded only during the approved check (4680.6844ms)
✔ a missing interpreter and repeated pytest item objects cannot create a false verification pass (2210.4558ms)
ℹ tests 13
ℹ suites 0
ℹ pass 13
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 41801.1199
@@ -0,0 +1,7 @@
PASS native parameters persist after reload and the narrow form remains usable without calling a model
PASS unsupported field is rejected visibly before any dispatch
PASS actual HTTP error shows raw fields and original wait, keeps secrets hidden and downloads SHA-verified original at both widths
PASS physical request ledger preserves the same error origin and offers the sealed actual request
PASS Responses failed event retains original identity and reported use at both widths; partial tool is never dispatched or retried
PASS explicit synthetic probe also shows the API diagnostic and labels its preview-only retention without inventing a Task artifact
{"success":true,"directory":"D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f","checks":6,"screenshots":10}
@@ -0,0 +1,6 @@
PASS native parameters persist after reload and the narrow form remains usable without calling a model
PASS unsupported field is rejected visibly before any dispatch
PASS actual HTTP error shows raw fields and original wait, keeps secrets hidden and downloads SHA-verified original at both widths
PASS physical request ledger preserves the same error origin and offers the sealed actual request
PASS Responses failed event retains original identity and reported use at both widths; partial tool is never dispatched or retried
{"success":true,"directory":"D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-f256eff7-e02f-4186-a9e6-d0f879fe82fe","checks":5,"screenshots":7}
@@ -0,0 +1 @@
Server bundle built.
@@ -0,0 +1,10 @@
✔ a directly started compiled host upgrades twice with the same PID, sessions, applied bytes and volatile provider (3988.2051ms)
✔ the running 0.1.0 process upgrades in place and retains its project, session, files and volatile model connection (8714.6336ms)
ℹ tests 2
ℹ suites 0
ℹ pass 2
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 12793.2125
@@ -0,0 +1,658 @@
{
"schema": 1,
"createdAt": "2026-11-01T15:00:03.778Z",
"version": "1.20.3",
"sourceCommit": "4bf4f7a8b271264420b7450c08a305529570d0dd",
"files": [
{
"path": "contained-ui/http-error-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\http-error-390.png",
"bytes": 42271,
"sha256": "62f389756b0dc1aa9195b62b4b38286ea1e5532f88650e042f282bab68aedea8"
},
{
"path": "contained-ui/http-error-desktop.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\http-error-desktop.png",
"bytes": 91112,
"sha256": "a26e4de08071acc816cf36d6364aa11cda1c6b46db3945749206c56b4f775ff5"
},
{
"path": "contained-ui/http-raw-desktop.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\http-raw-desktop.png",
"bytes": 91221,
"sha256": "772e60beb50e10dc591ea8c726d1554be0af23778e686d3646e5083eb7660570"
},
{
"path": "contained-ui/ledger-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\ledger-390.png",
"bytes": 53917,
"sha256": "0db258b881d9a7604b82c600fe264e176858a04e6d16866c63a5fc10587cfb5f"
},
{
"path": "contained-ui/native-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\native-390.png",
"bytes": 46719,
"sha256": "bf0c6db94447284972e44f196b5136a33a41e71510deec9d5d09202e9290537e"
},
{
"path": "contained-ui/native-desktop.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\native-desktop.png",
"bytes": 146953,
"sha256": "7eb38b0ccfb87074d9d6011d1b869142b9d0e54ff81ecc951b38845219e7a80f"
},
{
"path": "contained-ui/probe-error-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\probe-error-390.png",
"bytes": 49738,
"sha256": "d90c61454d8665df5e03ea837e2b5143e086c32d1c4ddf57aea38e67f476b263"
},
{
"path": "contained-ui/result.json",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d/result.json",
"bytes": 3283,
"sha256": "17fba29a7cf78255a64a093c97d1a25f9d9a874c9db59562bd69e5a1381024ae"
},
{
"path": "contained-ui/sse-error-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\sse-error-390.png",
"bytes": 49648,
"sha256": "2a56c56c95fcb934c61c4b6cbc3d16b7092bbfb4b48fb49e2a032f378fce9c66"
},
{
"path": "contained-ui/sse-error-desktop.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\sse-error-desktop.png",
"bytes": 94927,
"sha256": "9920c5e6318a14031d3f4b92204e3c1ed576dc784af40a76521ef9c090cce0b8"
},
{
"path": "contained-ui/sse-raw-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-10eb5e3f-33ed-4c59-b56f-7b442ff2b82d\\sse-raw-390.png",
"bytes": 49569,
"sha256": "8bdd5983369d1e2da56f66b81624c6d93025ab4649228b752e82cf725e81dc8c"
},
{
"path": "final-source-ui/http-error-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\http-error-390.png",
"bytes": 42256,
"sha256": "6272836b7ec03d0d5feef464a8023c6c62d0956454e45fae8f41798b57820d46"
},
{
"path": "final-source-ui/http-error-desktop.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\http-error-desktop.png",
"bytes": 90942,
"sha256": "67b3d629501ed4cb07917242374f776608938d03cc6d82826d4405ecad967bd2"
},
{
"path": "final-source-ui/http-raw-desktop.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\http-raw-desktop.png",
"bytes": 91117,
"sha256": "902d3900649c823d8dabe1c2860e82a37efafd455a80a5d15766d905c861ccc7"
},
{
"path": "final-source-ui/ledger-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\ledger-390.png",
"bytes": 53721,
"sha256": "13de4ab0fde40229b41b1e63cbb290adef467ea85310e01cea884f51c36bd4a9"
},
{
"path": "final-source-ui/native-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\native-390.png",
"bytes": 46719,
"sha256": "bf0c6db94447284972e44f196b5136a33a41e71510deec9d5d09202e9290537e"
},
{
"path": "final-source-ui/native-desktop.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\native-desktop.png",
"bytes": 146953,
"sha256": "7eb38b0ccfb87074d9d6011d1b869142b9d0e54ff81ecc951b38845219e7a80f"
},
{
"path": "final-source-ui/probe-error-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\probe-error-390.png",
"bytes": 50107,
"sha256": "7e8c19d768633df59cd301d5cc1051ad7652d89b59267be859fa4adbb73ffba2"
},
{
"path": "final-source-ui/result.json",
"source": ".qa/model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f/result.json",
"bytes": 3223,
"sha256": "78fd9aed265e92dcc60c15f4d726d4ab7ae2fc3b12cc9b56cba51fc9b4aafe00"
},
{
"path": "final-source-ui/sse-error-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\sse-error-390.png",
"bytes": 49521,
"sha256": "7bf3c42e0bcc7c76e0d5e7bdc73daa8a02bf2961146c8c550f1609fa0b759402"
},
{
"path": "final-source-ui/sse-error-desktop.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\sse-error-desktop.png",
"bytes": 94569,
"sha256": "db58adc605aaf1126cb1596e2cd28e8891d02c66612dc006c295a23556d646ce"
},
{
"path": "final-source-ui/sse-raw-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-c05ecba4-18c9-498e-8383-8bb99b0dc73f\\sse-raw-390.png",
"bytes": 49569,
"sha256": "8bdd5983369d1e2da56f66b81624c6d93025ab4649228b752e82cf725e81dc8c"
},
{
"path": "first-contained-ui/http-error-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\http-error-390.png",
"bytes": 42210,
"sha256": "08f5ced640518714ae99504b58b4986cf1ac53f3a6fc09653e908c2fa59eb54c"
},
{
"path": "first-contained-ui/http-error-desktop.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\http-error-desktop.png",
"bytes": 90865,
"sha256": "12ffc3b22da504d111919bf02bf287f9bb2ebca65b36117b843367e38345d9df"
},
{
"path": "first-contained-ui/http-raw-desktop.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\http-raw-desktop.png",
"bytes": 91150,
"sha256": "546c5382e99889e3b805add0bc97fe37442abbc13fd16577ad47a073d4f11ef0"
},
{
"path": "first-contained-ui/ledger-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\ledger-390.png",
"bytes": 53955,
"sha256": "a6e15fe4b02b1263609bfaeed0dc809445a28b5902d16c5378af6a7c0768d078"
},
{
"path": "first-contained-ui/native-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\native-390.png",
"bytes": 46719,
"sha256": "bf0c6db94447284972e44f196b5136a33a41e71510deec9d5d09202e9290537e"
},
{
"path": "first-contained-ui/native-desktop.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\native-desktop.png",
"bytes": 146953,
"sha256": "7eb38b0ccfb87074d9d6011d1b869142b9d0e54ff81ecc951b38845219e7a80f"
},
{
"path": "first-contained-ui/probe-error-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\probe-error-390.png",
"bytes": 49922,
"sha256": "90a122e70ae15238e3acd85040b7346a4dadc5bd9c722f14608647ff5448b5f0"
},
{
"path": "first-contained-ui/result.json",
"source": ".qa/model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860/result.json",
"bytes": 3283,
"sha256": "cb3302fe87741073933208cf0e9a49ec5d06ed2bfb08b2e85d050abbdcda0e8a"
},
{
"path": "first-contained-ui/sse-error-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\sse-error-390.png",
"bytes": 49658,
"sha256": "1dbb58980fba6d2282edf328c410d07799839deb49836aa5c8faee067c8dd0d6"
},
{
"path": "first-contained-ui/sse-error-desktop.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\sse-error-desktop.png",
"bytes": 95215,
"sha256": "f6d6c5e0804730dbab546c7ec57288ff32ba39c65b40c2251585ef7f4fd9d750"
},
{
"path": "first-contained-ui/sse-raw-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-cc04386a-194c-47be-91de-71d30319b860\\sse-raw-390.png",
"bytes": 49569,
"sha256": "8bdd5983369d1e2da56f66b81624c6d93025ab4649228b752e82cf725e81dc8c"
},
{
"path": "formal-ui-second/edge-1440-native-parameters.png",
"source": ".qa\\model-fidelity-formal-ui-second\\edge-1440-native-parameters.png",
"bytes": 152908,
"sha256": "22375864853ec335f115ce2df00dfe49bab6d07e4de0a9e83aa31bcada5dd95a"
},
{
"path": "formal-ui-second/edge-1440-original-session.png",
"source": ".qa\\model-fidelity-formal-ui-second\\edge-1440-original-session.png",
"bytes": 118948,
"sha256": "43c56823c6f791addd700272db67fc5c789fe3d951f11e19a6b7d6ee50e0f981"
},
{
"path": "formal-ui/edge-1440-native-parameters.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui\\edge-1440-native-parameters.png",
"bytes": 152908,
"sha256": "22375864853ec335f115ce2df00dfe49bab6d07e4de0a9e83aa31bcada5dd95a"
},
{
"path": "formal-ui/edge-1440-original-session.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui\\edge-1440-original-session.png",
"bytes": 118948,
"sha256": "43c56823c6f791addd700272db67fc5c789fe3d951f11e19a6b7d6ee50e0f981"
},
{
"path": "formal-ui/edge-390-native-parameters.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui\\edge-390-native-parameters.png",
"bytes": 41999,
"sha256": "3549e3ebedfc9545fde562a8ef949bb7d1e7ed903d64f6faad852316614e8832"
},
{
"path": "formal-ui/edge-390-original-session.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui\\edge-390-original-session.png",
"bytes": 45197,
"sha256": "e47bb6663a263e557dad7471f7a3384d1e06bc1279c4dce3592c764c38cd5aec"
},
{
"path": "formal-ui/electron-1280-native-parameters.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui\\electron-1280-native-parameters.png",
"bytes": 103709,
"sha256": "1c8c1325b91341b2493a960722c03a5fe48b6f4e726c2a5245213303309ff328"
},
{
"path": "formal-ui/electron-1280-original-session.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui\\electron-1280-original-session.png",
"bytes": 101882,
"sha256": "2d45acaa1a1c5e82de84c364ba0afcf8b6ee2e95e5be6c316fd2cf4d31c5254a"
},
{
"path": "formal-ui/electron-390-native-parameters.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui\\electron-390-native-parameters.png",
"bytes": 42680,
"sha256": "6a453ab9d24875414b40ca36b62a231da37aceb90331f9def1839f47286056b7"
},
{
"path": "formal-ui/electron-390-original-session.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui\\electron-390-original-session.png",
"bytes": 48912,
"sha256": "4c6092e8c2a582ba6bad5b206fa8e1cb640cb48d5a3bb85f8020ed41d763c402"
},
{
"path": "initial-ui/http-error-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-f256eff7-e02f-4186-a9e6-d0f879fe82fe\\http-error-390.png",
"bytes": 42489,
"sha256": "d0b383a4f4d6716f65d92fa072cd546c7488fc371910b528006743478bbea5a2"
},
{
"path": "initial-ui/http-error-desktop.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-f256eff7-e02f-4186-a9e6-d0f879fe82fe\\http-error-desktop.png",
"bytes": 91347,
"sha256": "1e5b870aa30622e7b40bea51a983f0199fe51d30e34fce2baec5b4e6917a6adc"
},
{
"path": "initial-ui/ledger-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-f256eff7-e02f-4186-a9e6-d0f879fe82fe\\ledger-390.png",
"bytes": 53792,
"sha256": "203a5ce85361c5c444f62a43c0277e24525e1c6f4ca19808c18bf4c0cd161b22"
},
{
"path": "initial-ui/native-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-f256eff7-e02f-4186-a9e6-d0f879fe82fe\\native-390.png",
"bytes": 46719,
"sha256": "bf0c6db94447284972e44f196b5136a33a41e71510deec9d5d09202e9290537e"
},
{
"path": "initial-ui/native-desktop.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-f256eff7-e02f-4186-a9e6-d0f879fe82fe\\native-desktop.png",
"bytes": 146953,
"sha256": "7eb38b0ccfb87074d9d6011d1b869142b9d0e54ff81ecc951b38845219e7a80f"
},
{
"path": "initial-ui/result.json",
"source": ".qa/model-fidelity-ui-f256eff7-e02f-4186-a9e6-d0f879fe82fe/result.json",
"bytes": 2378,
"sha256": "17a8b2c7ed646cf7577dae66c283633a31238a8ec19ddfd36e3fecaf8111b6d7"
},
{
"path": "initial-ui/sse-error-390.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-f256eff7-e02f-4186-a9e6-d0f879fe82fe\\sse-error-390.png",
"bytes": 48281,
"sha256": "20d3270db03ac8edce6b6af2db48346f39150ab7c607a33d19936940081b352c"
},
{
"path": "initial-ui/sse-error-desktop.png",
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-ui-f256eff7-e02f-4186-a9e6-d0f879fe82fe\\sse-error-desktop.png",
"bytes": 90810,
"sha256": "c849776a7a49b4fd2fc29b632ba4719d6004abe12676f8c3b1a39057195125e7"
},
{
"path": "logs/model-fidelity-bind-package.txt",
"source": ".qa/model-fidelity-bind-package.log",
"bytes": 165,
"sha256": "181bae890b39da589d7fb74539feab5d45ab82b7402a59158394d69d80cbcacb"
},
{
"path": "logs/model-fidelity-build-final.txt",
"source": ".qa/model-fidelity-build-final.log",
"bytes": 838,
"sha256": "aa98267791b73d08a6a2bfd82a3f59ad2267a16c74783b92c3e7a5a578d98ec8"
},
{
"path": "logs/model-fidelity-build.txt",
"source": ".qa/model-fidelity-build.log",
"bytes": 838,
"sha256": "1624a9c847011e45756adf907ab62f6b04989bba1ae79a59aac317c496ac4bb0"
},
{
"path": "logs/model-fidelity-contained-ui.txt",
"source": ".qa/model-fidelity-contained-ui.log",
"bytes": 828,
"sha256": "5781da791973d90353db185f64ead8e49e28428736e04a38dc572da492e5039f"
},
{
"path": "logs/model-fidelity-contracts.tap",
"source": ".qa/model-fidelity-contracts.tap",
"bytes": 6100,
"sha256": "c79f2a7d0d09cd162fd0a5c3b4c21c72a0731e25c71033a6a4b4ee28451236ce"
},
{
"path": "logs/model-fidelity-final-related.tap",
"source": ".qa/model-fidelity-final-related.tap",
"bytes": 12085,
"sha256": "bb6b4e178058326e30414c4d3831a75b4501279461dd3b5f707238fb1ad1952b"
},
{
"path": "logs/model-fidelity-first-bind-package.txt",
"source": ".qa/model-fidelity-first-bind-package.log",
"bytes": 165,
"sha256": "eff5b1a64f58a13b09a257b65bbe7146bf25340f2058fbc8c80fae0b346af17b"
},
{
"path": "logs/model-fidelity-first-contained-ui.txt",
"source": ".qa/model-fidelity-first-contained-ui.log",
"bytes": 828,
"sha256": "0d0fdd6fc5ce2ca5bd8048f9cfd0d63c1b0cb782a051842c2dd257b93bdd8628"
},
{
"path": "logs/model-fidelity-first-owned-upgrade.txt",
"source": ".qa/model-fidelity-first-owned-upgrade.log",
"bytes": 892,
"sha256": "be1b0eba4639a6b496a077cbccc1b6849ef2f377e9869b01c60e59923f4ff01f"
},
{
"path": "logs/model-fidelity-first-package.txt",
"source": ".qa/model-fidelity-first-package.log",
"bytes": 2946,
"sha256": "d0cbf9e3138fa69ee140aceaabeed671ea26ea3c939db75c43eb6a1f3bbdce0a"
},
{
"path": "logs/model-fidelity-formal-activate.txt",
"source": ".qa/model-fidelity-formal-activate.log",
"bytes": 621,
"sha256": "87093fcfcb1d0986064c1b27af72de9e390ce1533578f34d461159e1668ce93f"
},
{
"path": "logs/model-fidelity-formal-stage.txt",
"source": ".qa/model-fidelity-formal-stage.log",
"bytes": 83,
"sha256": "be2ad129332966b4a395af6ea04b32f5f59bf943ea93f90b85578016da609287"
},
{
"path": "logs/model-fidelity-formal-ui-first.txt",
"source": ".qa/model-fidelity-formal-ui-first.log",
"bytes": 560,
"sha256": "d8efee5fd19f1b3db45291d3ea35edc924cb9690b224ebbc9588f910be76ca1a"
},
{
"path": "logs/model-fidelity-formal-ui-second.txt",
"source": ".qa/model-fidelity-formal-ui-second.log",
"bytes": 1320,
"sha256": "d5962bef1dfddd24ccbb5311d30ee2c30489ec5070ea37f7726c9422e4b15811"
},
{
"path": "logs/model-fidelity-formal-ui.txt",
"source": ".qa/model-fidelity-formal-ui.log",
"bytes": 46,
"sha256": "21b53c9bf795411fd51cc09508d27a87776e760c90f558e115784a1f22b5b188"
},
{
"path": "logs/model-fidelity-formal-upgrade.txt",
"source": ".qa/model-fidelity-formal-upgrade.log",
"bytes": 87,
"sha256": "b6503734fdd549638f69ae20485c9cf6b82568bf67ea6d49ee2eae4e6b1b11a2"
},
{
"path": "logs/model-fidelity-formal-verify.txt",
"source": ".qa/model-fidelity-formal-verify.log",
"bytes": 701,
"sha256": "ae3b8c097d37287e958b8f4681db08ef3720e7579a76618f3feb97a413fb9402"
},
{
"path": "logs/model-fidelity-full.tap",
"source": ".qa/model-fidelity-full.tap",
"bytes": 79855,
"sha256": "c34cfa4fabb8fc04dae986a394430be5a22655ab36bf574e1bfad9365df4fdd7"
},
{
"path": "logs/model-fidelity-initial.tap",
"source": ".qa/model-fidelity-initial.tap",
"bytes": 2776,
"sha256": "bfc3e5d8585f495c703b55edbdb10f2e17e88a640d4201cc473b0f466599a978"
},
{
"path": "logs/model-fidelity-install-final.tap",
"source": ".qa/model-fidelity-install-final.tap",
"bytes": 1380,
"sha256": "a6bae06df33d2cab4d6a5ad3b91ac4aca731fa598692ed50cff5e0e6993e3744"
},
{
"path": "logs/model-fidelity-native-final.tap",
"source": ".qa/model-fidelity-native-final.tap",
"bytes": 2181,
"sha256": "69c0e03bcdcc841d5fba83079efbf659ed7624fe8b7c250ef9959f4c55d49322"
},
{
"path": "logs/model-fidelity-owned-upgrade.txt",
"source": ".qa/model-fidelity-owned-upgrade.log",
"bytes": 142,
"sha256": "a31a17afee9aeb54f7823f448fe2ce6187e255fa6f971b0b0595859fbffecdd8"
},
{
"path": "logs/model-fidelity-package.txt",
"source": ".qa/model-fidelity-package.log",
"bytes": 2946,
"sha256": "52901d34f2e8e73a4400b3e0807ee419e3945bda363a1133cf9f1f97735ad3e1"
},
{
"path": "logs/model-fidelity-python-final.tap",
"source": ".qa/model-fidelity-python-final.tap",
"bytes": 1656,
"sha256": "15631df1849f0a2585b2f4b60e31d1b1e92831a237597c82b453443d605e22c5"
},
{
"path": "logs/model-fidelity-ui-final.txt",
"source": ".qa/model-fidelity-ui-final.log",
"bytes": 828,
"sha256": "153898bfb7427149f213b2a17c0e992c2e11acdcfd2ab55baa3475e3beb5f189"
},
{
"path": "logs/model-fidelity-ui.txt",
"source": ".qa/model-fidelity-ui.log",
"bytes": 694,
"sha256": "3905f1a241436c31d5b4a017ee9f9ce89d2df3001ed572483b9575454c352dc6"
},
{
"path": "logs/model-fidelity-upgrade-build.txt",
"source": ".qa/model-fidelity-upgrade-build.log",
"bytes": 22,
"sha256": "eb0215f4d46d26cf809f82eb6fd9b77f450ea30825c907f180ef608a40706c8c"
},
{
"path": "logs/model-fidelity-upgrade-fix.tap",
"source": ".qa/model-fidelity-upgrade-fix.tap",
"bytes": 388,
"sha256": "f435a74c37b0e58f48d74e7152660bad841098549f1a1415e31c6328b3384729"
},
{
"path": "receipts/model-fidelity-activity-proof.json",
"source": ".qa/model-fidelity-activity-proof.json",
"bytes": 1519,
"sha256": "31b78e83aeddfe19aaeb167eb5ed01c4e99457f3afb55505423c79e9037fd99e"
},
{
"path": "receipts/model-fidelity-first-owned-upgrade.json",
"source": ".qa/model-fidelity-first-owned-upgrade.json",
"bytes": 445,
"sha256": "39d4eeb1dcabd55da6da62700676b76bcd17fc226ba37ecfa7bed33d29632f6d"
},
{
"path": "receipts/model-fidelity-first-package-binding.json",
"source": ".qa/model-fidelity-first-package-binding.json",
"bytes": 98527,
"sha256": "d62cff0cf25f8957965686394573be6ab9522ad4ac86c470d522c07f5f7928f6"
},
{
"path": "receipts/model-fidelity-first-release.json",
"source": ".qa/model-fidelity-first-release.json",
"bytes": 2875,
"sha256": "71988d7eee15069cac5c18e4a0ee28ef8ac1175109fe2efd59a024bec0f4f7c5"
},
{
"path": "receipts/model-fidelity-formal-activate-rejected.json",
"source": ".qa/model-fidelity-formal-activate.json",
"bytes": 1434,
"sha256": "868c8df6e920e1132cec6f8ebde51d8d16f200fb63634cc01a720281e7872796"
},
{
"path": "receipts/model-fidelity-formal-observed-activation.json",
"source": ".qa/model-fidelity-formal-observed-activation.json",
"bytes": 1264,
"sha256": "793585b4cd13517910f2ec4e22db6631a82d542f1a764a4b74f3f3ef1fa67bd8"
},
{
"path": "receipts/model-fidelity-formal-stage.json",
"source": ".qa/model-fidelity-formal-stage.json",
"bytes": 1438,
"sha256": "d49030ce112ca410347870781e477a3da89d4c5371b08e9b56c0e0f035457c33"
},
{
"path": "receipts/model-fidelity-formal-ui-second.json",
"source": ".qa/model-fidelity-formal-ui-second.json",
"bytes": 2090,
"sha256": "0642f44ca4a8a1a70b6ea44dd2350b829ae09588b8391731bb6c6a0e6219a0f2"
},
{
"path": "receipts/model-fidelity-formal-ui.json",
"source": ".qa/model-fidelity-formal-ui.json",
"bytes": 3727,
"sha256": "5b4ff1c32f6aadedc4a9adf56936a6ddf7cababecbda2a5f24aa2aba868c877d"
},
{
"path": "receipts/model-fidelity-formal-upgrade.json",
"source": ".qa/model-fidelity-formal-upgrade.json",
"bytes": 1467,
"sha256": "2ab805f9732256cdb765e69b4827ba3cc62df6927a66e3df27fae701db175498"
},
{
"path": "receipts/model-fidelity-formal-verify.json",
"source": ".qa/model-fidelity-formal-verify.json",
"bytes": 3669,
"sha256": "1cb2382ba54ff32c83a17a393f244d8c9535069fc795714aa14ec3d35d3c4131"
},
{
"path": "receipts/model-fidelity-owned-upgrade.json",
"source": ".qa/model-fidelity-owned-upgrade.json",
"bytes": 21494,
"sha256": "59d3d5481d22ac06b25b45cd4759289c1f566d365427ce96a08589c5b4e6eac9"
},
{
"path": "receipts/model-fidelity-package-binding.json",
"source": ".qa/model-fidelity-package-binding.json",
"bytes": 98527,
"sha256": "14d7399ed6cedccbb920deab5c96d6409230fef2888a3ab17dec22b927f95e4d"
},
{
"path": "receipts/model-fidelity-package-validation.json",
"source": ".qa/model-fidelity-package-validation.json",
"bytes": 982,
"sha256": "d58b8f0b41b532e1b14325dce2bfaeebab46c5b6ed48c48f0f893f26020d5c43"
},
{
"path": "retained-summary.json",
"source": ".qa/model-fidelity-retained-summary.json",
"bytes": 4940,
"sha256": "2cc9f08b3662eee335cfb7f9e7a1a7d659fee1f8b84891cc681e4894923c713f"
},
{
"path": "verifiers/desktop-f-state.mjs",
"source": ".qa/desktop-f-state.mjs",
"bytes": 5932,
"sha256": "6cf41ebd2ce55fe6fe3175948e9b964fac99975aca3132420a83b1d122e08b0b"
},
{
"path": "verifiers/model-fidelity-bind-package.mjs",
"source": ".qa/model-fidelity-bind-package.mjs",
"bytes": 4186,
"sha256": "de69b19f38d69cab7f9cccc6ba087bf2e53b2821db02173da342fb0f64419efd"
},
{
"path": "verifiers/model-fidelity-final-audit.mjs",
"source": ".qa/model-fidelity-final-audit.mjs",
"bytes": 4354,
"sha256": "4e8b82086de84c84819eb45f8e1f51603b55f98338a893e33fc3328b1791006b"
},
{
"path": "verifiers/model-fidelity-finalize-docs.mjs",
"source": ".qa/model-fidelity-finalize-docs.mjs",
"bytes": 13821,
"sha256": "e1f9942c4ee6b4a3411254f60ddf6bb36f8485c38c895b06bc534dc6f6af1d34"
},
{
"path": "verifiers/model-fidelity-first-bind-package.mjs",
"source": ".qa/model-fidelity-first-bind-package.mjs",
"bytes": 4186,
"sha256": "3d3f50251d9266d4cb3ff0ac318329ab21bdcb091ed7a31e8d3383a7df24719a"
},
{
"path": "verifiers/model-fidelity-formal-ui-first.mjs",
"source": ".qa/model-fidelity-formal-ui-first.mjs",
"bytes": 6582,
"sha256": "164821be44d523d858e8767ded09cc2d0fa0a604d07143f56176d10f222ca022"
},
{
"path": "verifiers/model-fidelity-formal-ui-second.mjs",
"source": ".qa/model-fidelity-formal-ui-second.mjs",
"bytes": 6703,
"sha256": "4b8275bfb52670eb0d1f7ad86daf6e967537099932a6466d9c2904af1ddd3bb6"
},
{
"path": "verifiers/model-fidelity-formal-ui.mjs",
"source": ".qa/model-fidelity-formal-ui.mjs",
"bytes": 6704,
"sha256": "eac3899d451b29da29d0017f5bf656aad81e5d15def0fac9162af1afc67104a2"
},
{
"path": "verifiers/model-fidelity-formal.mjs",
"source": ".qa/model-fidelity-formal.mjs",
"bytes": 9356,
"sha256": "cb43b0590a1def00b0b135c9494dc9415371b6a1256cfb6f5761d4f75849eee9"
},
{
"path": "verifiers/model-fidelity-owned-upgrade.mjs",
"source": ".qa/model-fidelity-owned-upgrade.mjs",
"bytes": 2597,
"sha256": "a725daf079b4cf4f4bd3b45614765f83481bd27b09f693fca19114c70c36810f"
},
{
"path": "verifiers/model-fidelity-reconcile.mjs",
"source": ".qa/model-fidelity-reconcile.mjs",
"bytes": 2114,
"sha256": "98619c13688f3989d631b2ef51c2f6890b752c7c40e5e4736cea8cc8d43feed3"
},
{
"path": "verifiers/model-fidelity-resume-verify.mjs",
"source": ".qa/model-fidelity-resume-verify.mjs",
"bytes": 11585,
"sha256": "9aa0174bf6671a6d973a7eda7d3d1ae3c8f5382fe964135d823aef656740c936"
},
{
"path": "verifiers/model-fidelity-seal.mjs",
"source": ".qa/model-fidelity-seal.mjs",
"bytes": 10216,
"sha256": "4d5c2dd31a226974636fa1cef34510edccb4dd265e2a03aec544129822ecc8dd"
},
{
"path": "verifiers/model-fidelity-validate-package.mjs",
"source": ".qa/model-fidelity-validate-package.mjs",
"bytes": 2394,
"sha256": "02a05a03723089172d77432cd511174b9a76431fe1da9a1baa42effff5ebba9c"
}
],
"realModelRequests": 0,
"scope": "Exact local controlled verification logs, screenshots and deployment receipts; excludes private SQLite, full original state/export inventories and credentials. Verifier sources retain their original .qa relative imports."
}
@@ -0,0 +1,34 @@
{
"success": true,
"sessionId": "90d02c6d-e45a-405d-891d-13f012c9f0f5",
"taskId": "f4ef3093-5a53-4c7e-8b50-4b3a80b8cb01",
"runId": "4330ba40-3f13-47d4-93c9-4e466096814f",
"beforeSessions": 31,
"sessions": 32,
"additions": {
"sessions": 1,
"records": 16,
"events": 57,
"outbox": 57,
"artifacts": 9,
"attempts": 2
},
"originalStateRetained": {
"originalRows": "exact except modelAttempts appended two owned user attempts and events sequence advanced",
"originalExports": "all 31 exact",
"originalArtifactBytes": "all original exact; nine new same-session artifacts retained",
"originalModelCenter": "all profiles/routes/credentials/original attempts exact; two user attempts and their connection health retained",
"stableCurrentRowsExportsArtifactsAndModelCenter": {
"originalRows": "exact",
"originalExports": "exact",
"originalArtifactBytes": "exact",
"originalModelCenter": "exact"
}
},
"stableBaseline": ".qa/model-fidelity-formal-stable-before.json",
"userRunStartedAt": "2026-11-01T14:50:14.276Z",
"userRunFinishedAt": "2026-11-01T14:51:46.915Z",
"realModelRequestsInitiatedByValidation": 0,
"userObservedModelRequests": 2,
"scope": "Strict original-row/export/artifact subset comparison plus explicit same-session additions, exact old attempt metadata and sole used connection health, followed by unmodified full original-state comparison at fresh stable baseline. No install, upgrade, model or task replay."
}
@@ -0,0 +1,5 @@
{
"success": false,
"error": "Error: This migration accepts a supported existing Workbench.\n at upgradeRunning (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/scripts/upgrade-running.mjs:67:324)\n at process.processTicksAndRejections (node:internal/process/task_queues:104:5)\n at async file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/.qa/model-fidelity-owned-upgrade.mjs:17:20",
"realModelRequests": 0
}
@@ -0,0 +1,70 @@
{
"portableRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-win-x64-20261101140547938",
"sourceRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-source-20261101140547938",
"portable": {
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-win-x64-20261101140547938.zip",
"bytes": 629123043,
"sha256": "419dfe81ab32e3dc0abc23f2714e2f603bc06c4fb1941f6db4dca050204e8b19"
},
"source": {
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-source-20261101140547938.zip",
"bytes": 629305353,
"sha256": "8cfda0726019d9bb119cc4b9be0aff7a4cb4c76d61106791060315ac920ad56a"
},
"installer": {
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-win-x64-20261101140547938-Setup.exe",
"bytes": 588262697,
"sha256": "5cb96077abc5b6a4339475228804761df65907004c1904f221827f5c71623f46",
"compiler": {
"version": "6.3.1",
"sha256": "3b22d36fc7e7c421d8caf780008ecdb42b14b3f0371b9bafe3cad64e6648d9cf"
},
"source": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\installer-build-18e26917-bee3-4b41-bdc4-09f7a01ed915\\installer.iss",
"version": "1.20.3",
"candidateDirectory": "Agent-Workbench-1.20.3-win-x64-20261101140547938",
"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": 19638,
"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 +1,19 @@
{
"success": false,
"phase": "activate",
"error": "AssertionError [ERR_ASSERTION]: sessionsSha256\n+ actual - expected\n\n+ '11c2bca842fa257402b23224957a23d11441b86aa2b94e6a4344a0e8960c579e'\n- 'cdecf3e1f8b3ba36d168e17394ac509ec8394b8adc7748e43b7e21bf16482168'\n\n at compareOriginalState (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/.qa/desktop-f-state.mjs:47:116)\n at file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/.qa/model-fidelity-formal.mjs:69:69\n at process.processTicksAndRejections (node:internal/process/task_queues:104:5)",
"startedAt": "2026-11-01T14:41:30.586Z",
"finishedAt": "2026-11-01T14:53:40.932Z",
"installRoot": "C:/Users/wangy/AppData/Local/Programs/AgentWorkbench",
"dataRoot": "C:\\Users\\wangy\\AppData\\Local\\Packages\\OpenAI.Codex_2p2nqsd0c76g0\\LocalCache\\Local\\AgentWorkbench",
"candidate": "C:\\Users\\wangy\\AppData\\Local\\Programs\\AgentWorkbench\\versions\\Agent-Workbench-1.20.3-win-x64-20261101141644779",
"setupSha256": "c30c362c58641749b884d47c4ed906063074024eab19e8e494cd81afa91c0141",
"originalProcess": {
"pid": 36552,
"parentPid": 32968,
"createdAt": "2026-10-31T06:30:34.359Z",
"executable": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.14.0-win-x64-20261030235440461\\runtime\\node.exe",
"userSid": "S-1-5-21-3906741337-925321518-3767092861-1001"
},
"realModelRequests": 0
}
@@ -0,0 +1,26 @@
{
"success": true,
"phase": "observed-completed-activation",
"actualActivationRepeated": false,
"originalRejectedOuterVerification": ".qa/model-fidelity-formal-activate.json",
"originalStartedAt": "2026-11-01T14:41:30.586Z",
"originalFinishedAt": "2026-11-01T14:53:40.932Z",
"productOperation": {
"path": "C:\\Users\\wangy\\AppData\\Local\\Programs\\AgentWorkbench\\.installation\\operation-446dcfc6-dd58-4bea-a22b-6ff00adf58ba.json",
"sha256": "9b7ed690ee17c66c6ceac83199da5b8269b546410bf815ac3c31d9497ee844ae",
"state": "activated"
},
"pointerSha256": "20ed12b6c97053b2201b4f10953da75e24a7e95a6b52a99668ed07a205638311",
"retained": {
"originalRows": "exact except modelAttempts appended two owned user attempts and events sequence advanced",
"originalExports": "all 31 exact",
"originalArtifactBytes": "all original exact; nine new same-session artifacts retained",
"originalModelCenter": "all profiles/routes/credentials/original attempts exact; two user attempts and their connection health retained",
"stableCurrentRowsExportsArtifactsAndModelCenter": {
"originalRows": "exact",
"originalExports": "exact",
"originalArtifactBytes": "exact",
"originalModelCenter": "exact"
}
}
}
@@ -0,0 +1,29 @@
{
"success": true,
"phase": "staged",
"version": "1.20.3",
"candidateIdentity": "a9dff9f51d7b2d1ebbaba7dd01ea46e5995847ec7aa412d59aadf84efd704c6b",
"manifestSha256": "03ebae3f37134c88ca9cab399bca983cdcc303928b8490a8c0d597cf477a23e7",
"retained": {
"originalRows": "exact",
"originalExports": "exact",
"originalArtifactBytes": "exact",
"originalModelCenter": "exact"
},
"setupLog": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-setup.log",
"setupLogSha256": "3426a70eb91d1e4ef2c42cfbc8a2e0e3ab4ac7034db498498f86e7f4322bd447",
"startedAt": "2026-11-01T14:26:14.259Z",
"finishedAt": "2026-11-01T14:40:49.100Z",
"installRoot": "C:/Users/wangy/AppData/Local/Programs/AgentWorkbench",
"dataRoot": "C:\\Users\\wangy\\AppData\\Local\\Packages\\OpenAI.Codex_2p2nqsd0c76g0\\LocalCache\\Local\\AgentWorkbench",
"candidate": "C:\\Users\\wangy\\AppData\\Local\\Programs\\AgentWorkbench\\versions\\Agent-Workbench-1.20.3-win-x64-20261101141644779",
"setupSha256": "c30c362c58641749b884d47c4ed906063074024eab19e8e494cd81afa91c0141",
"originalProcess": {
"pid": 36552,
"parentPid": 32968,
"createdAt": "2026-10-31T06:30:34.359Z",
"executable": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.14.0-win-x64-20261030235440461\\runtime\\node.exe",
"userSid": "S-1-5-21-3906741337-925321518-3767092861-1001"
},
"realModelRequests": 0
}
@@ -0,0 +1,28 @@
{
"success": false,
"results": [
{
"label": "edge-1440",
"width": 1440,
"height": 1000,
"originalStatus": "completed",
"originalSessionPreserved": true,
"nativeSettingsMatchPublicConnection": true,
"saveFullyReachable": true,
"submitted": false
}
],
"screenshots": [
{
"file": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui\\edge-1440-original-session.png",
"sha256": "43c56823c6f791addd700272db67fc5c789fe3d951f11e19a6b7d6ee50e0f981"
},
{
"file": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui\\edge-1440-native-parameters.png",
"sha256": "22375864853ec335f115ce2df00dfe49bab6d07e4de0a9e83aa31bcada5dd95a"
}
],
"errors": [],
"blockedWrites": [],
"error": "Error: expect(locator).toBeVisible() failed\n\nLocator: getByText('创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D', { exact: true }).first()\nExpected: visible\nReceived: hidden\nTimeout: 5000ms\n\nCall log:\n - Expect \"toBeVisible\" getByText('创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D', { exact: true }).first() with timeout 5000ms\n - waiting for getByText('创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D', { exact: true }).first()\n 14 × locator resolved to <span>创建一个HTML,内容是SVG绘制一个鹈鹕骑自行车的2D</span>\n - unexpected value \"hidden\"\n\n at captureRawStack (D:\\Ai-project\\REA逆向\\Agent-Workbench\\node_modules\\playwright-core\\lib\\coreBundle.js:9282:17)\n at callMatcherAsStep (D:\\Ai-project\\REA逆向\\Agent-Workbench\\node_modules\\playwright\\lib\\matchers\\expect.js:13316:57)\n at Object.toBeVisible (D:\\Ai-project\\REA逆向\\Agent-Workbench\\node_modules\\playwright\\lib\\matchers\\expect.js:13308:23)\n at inspect (D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui.mjs:30:81)\n at async file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/.qa/model-fidelity-formal-ui.mjs:52:5"
}
@@ -0,0 +1,94 @@
{
"success": true,
"version": "1.20.3",
"pid": 36552,
"candidate": "C:\\Users\\wangy\\AppData\\Local\\Programs\\AgentWorkbench\\versions\\Agent-Workbench-1.20.3-win-x64-20261101141644779",
"results": [
{
"label": "edge-1440",
"width": 1440,
"height": 1000,
"originalStatus": "completed",
"originalSessionPreserved": true,
"nativeSettingsMatchPublicConnection": true,
"saveFullyReachable": true,
"submitted": false
},
{
"label": "edge-390",
"width": 390,
"height": 844,
"originalStatus": "completed",
"originalSessionPreserved": true,
"nativeSettingsMatchPublicConnection": true,
"saveFullyReachable": true,
"submitted": false
},
{
"label": "electron-1280",
"width": 1280,
"height": 900,
"originalStatus": "completed",
"originalSessionPreserved": true,
"nativeSettingsMatchPublicConnection": true,
"saveFullyReachable": true,
"submitted": false
},
{
"label": "electron-390",
"width": 390,
"height": 844,
"originalStatus": "completed",
"originalSessionPreserved": true,
"nativeSettingsMatchPublicConnection": true,
"saveFullyReachable": true,
"submitted": false
}
],
"screenshots": [
{
"file": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui\\edge-1440-original-session.png",
"sha256": "43c56823c6f791addd700272db67fc5c789fe3d951f11e19a6b7d6ee50e0f981"
},
{
"file": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui\\edge-1440-native-parameters.png",
"sha256": "22375864853ec335f115ce2df00dfe49bab6d07e4de0a9e83aa31bcada5dd95a"
},
{
"file": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui\\edge-390-original-session.png",
"sha256": "e47bb6663a263e557dad7471f7a3384d1e06bc1279c4dce3592c764c38cd5aec"
},
{
"file": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui\\edge-390-native-parameters.png",
"sha256": "3549e3ebedfc9545fde562a8ef949bb7d1e7ed903d64f6faad852316614e8832"
},
{
"file": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui\\electron-1280-original-session.png",
"sha256": "2d45acaa1a1c5e82de84c364ba0afcf8b6ee2e95e5be6c316fd2cf4d31c5254a"
},
{
"file": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui\\electron-1280-native-parameters.png",
"sha256": "1c8c1325b91341b2493a960722c03a5fe48b6f4e726c2a5245213303309ff328"
},
{
"file": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui\\electron-390-original-session.png",
"sha256": "4c6092e8c2a582ba6bad5b206fa8e1cb640cb48d5a3bb85f8020ed41d763c402"
},
{
"file": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\.qa\\model-fidelity-formal-ui\\electron-390-native-parameters.png",
"sha256": "6a453ab9d24875414b40ca36b62a231da37aceb90331f9def1839f47286056b7"
}
],
"errors": [],
"blockedWrites": [],
"retained": {
"originalRows": "exact",
"originalExports": "exact",
"originalArtifactBytes": "exact",
"originalModelCenter": "exact"
},
"originalSession": "22c90b01-6b82-43d9-b2b6-88a4b245c781",
"actualInstalledElectron": true,
"realModelRequests": 0,
"scope": "Read-only formal Edge and visible installed Electron at two viewports; actual original session/current status and native parameters/save reachability. Original historical session was already completed at deployment baseline; no invented failure display. No settings saved, models called or tasks retried."
}
@@ -0,0 +1,37 @@
{
"success": true,
"phase": "upgraded",
"version": "1.20.3",
"upgraded": {
"version": "1.20.3",
"pid": 36552,
"hasKey": true,
"projects": 17,
"sessions": 31,
"appDirectory": "C:\\Users\\wangy\\AppData\\Local\\Programs\\AgentWorkbench\\versions\\Agent-Workbench-1.20.3-win-x64-20261101141644779",
"verified": true,
"connectionPreserved": true,
"url": "http://127.0.0.1:4318"
},
"retained": {
"originalRows": "exact",
"originalExports": "exact",
"originalArtifactBytes": "exact",
"originalModelCenter": "exact"
},
"currentProviderPreserved": true,
"startedAt": "2026-11-01T14:41:03.688Z",
"finishedAt": "2026-11-01T14:41:12.796Z",
"installRoot": "C:/Users/wangy/AppData/Local/Programs/AgentWorkbench",
"dataRoot": "C:\\Users\\wangy\\AppData\\Local\\Packages\\OpenAI.Codex_2p2nqsd0c76g0\\LocalCache\\Local\\AgentWorkbench",
"candidate": "C:\\Users\\wangy\\AppData\\Local\\Programs\\AgentWorkbench\\versions\\Agent-Workbench-1.20.3-win-x64-20261101141644779",
"setupSha256": "c30c362c58641749b884d47c4ed906063074024eab19e8e494cd81afa91c0141",
"originalProcess": {
"pid": 36552,
"parentPid": 32968,
"createdAt": "2026-10-31T06:30:34.359Z",
"executable": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.14.0-win-x64-20261030235440461\\runtime\\node.exe",
"userSid": "S-1-5-21-3906741337-925321518-3767092861-1001"
},
"realModelRequests": 0
}
@@ -0,0 +1,81 @@
{
"success": true,
"phase": "verify-after-observed-user-activity",
"startedAt": "2026-11-01T14:56:45.447Z",
"finishedAt": "2026-11-01T14:56:56.346Z",
"installRoot": "C:/Users/wangy/AppData/Local/Programs/AgentWorkbench",
"dataRoot": "C:\\Users\\wangy\\AppData\\Local\\Packages\\OpenAI.Codex_2p2nqsd0c76g0\\LocalCache\\Local\\AgentWorkbench",
"candidate": "C:\\Users\\wangy\\AppData\\Local\\Programs\\AgentWorkbench\\versions\\Agent-Workbench-1.20.3-win-x64-20261101141644779",
"identity": {
"schema": 1,
"protocol": "agent-workbench-host-v1",
"app": "agent-workbench",
"version": "1.20.3",
"instanceId": "4c910dd5-82cb-4c8a-92f6-328338ea0662",
"pid": 36552,
"createdAt": "2026-10-31T06:30:34.359Z",
"userSid": "S-1-5-21-3906741337-925321518-3767092861-1001",
"ownerId": "b3ae90d3-9aa1-4b40-bd96-a53f01dbb502",
"executable": {
"path": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.14.0-win-x64-20261030235440461\\runtime\\node.exe",
"bytes": 93915464,
"sha256": "d702daa815eabbff763c7d3808e235cf031cf2ca70211b56847bc72f51e240e5"
},
"appRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.14.0-win-x64-20261030235440461",
"runtimeRoot": "C:\\Users\\wangy\\AppData\\Local\\Programs\\AgentWorkbench\\versions\\Agent-Workbench-1.20.3-win-x64-20261101141644779",
"runtimeIdentity": "a9dff9f51d7b2d1ebbaba7dd01ea46e5995847ec7aa412d59aadf84efd704c6b",
"dataRoot": "C:\\Users\\wangy\\AppData\\Local\\Packages\\OpenAI.Codex_2p2nqsd0c76g0\\LocalCache\\Local\\AgentWorkbench",
"port": 4318
},
"pointer": {
"schema": 1,
"app": "agent-workbench-installation",
"installId": "74c11bdd-4895-459b-8403-097f79980b41",
"revision": 10,
"dataRoot": "C:\\Users\\wangy\\AppData\\Local\\Packages\\OpenAI.Codex_2p2nqsd0c76g0\\LocalCache\\Local\\AgentWorkbench",
"port": 4318,
"active": {
"directory": "Agent-Workbench-1.20.3-win-x64-20261101141644779",
"version": "1.20.3",
"manifestSha256": "03ebae3f37134c88ca9cab399bca983cdcc303928b8490a8c0d597cf477a23e7",
"runtimeIdentity": "a9dff9f51d7b2d1ebbaba7dd01ea46e5995847ec7aa412d59aadf84efd704c6b"
},
"previous": {
"directory": "Agent-Workbench-1.20.2-win-x64-20261101124548744",
"version": "1.20.2",
"manifestSha256": "0fdfd44cecb46ebc4fdfc4aac24bec9587943016f744f8d7dd6d493adc418b8d",
"runtimeIdentity": "611c7ef2ffd984dde065dbfc77b7d54c666f48bcfb26557995a67f9f9d732bd4"
},
"activatedAt": "2026-11-01T14:53:34.487Z"
},
"retained": {
"originalRows": "exact except modelAttempts appended two owned user attempts and events sequence advanced",
"originalExports": "all 31 exact",
"originalArtifactBytes": "all original exact; nine new same-session artifacts retained",
"originalModelCenter": "all profiles/routes/credentials/original attempts exact; two user attempts and their connection health retained",
"stableCurrentRowsExportsArtifactsAndModelCenter": {
"originalRows": "exact",
"originalExports": "exact",
"originalArtifactBytes": "exact",
"originalModelCenter": "exact"
}
},
"currentProviderPreserved": true,
"activityProof": ".qa/model-fidelity-activity-proof.json",
"observedActivation": ".qa/model-fidelity-formal-observed-activation.json",
"companion": {
"enabled": false,
"devices": 0,
"certificate": null
},
"counts": {
"projects": 17,
"visibleSessions": 22,
"sessions": 32,
"records": 3239,
"events": 6229,
"artifacts": 1780
},
"inspectorClosed": true,
"realModelRequests": 0
}
@@ -0,0 +1,610 @@
{
"success": true,
"from": "1.20.2",
"version": "1.20.3",
"retained": {
"originalRows": "exact",
"originalExports": "exact",
"originalArtifactBytes": "exact",
"originalModelCenter": "exact"
},
"upgraded": {
"version": "1.20.3",
"pid": 38928,
"hasKey": false,
"projects": 1,
"sessions": 1,
"appDirectory": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-win-x64-20261101141644779",
"verified": true,
"connectionPreserved": true,
"url": "http://127.0.0.1:61608"
},
"before": {
"observedAt": "2026-11-01T14:25:51.333Z",
"version": "1.20.2",
"pid": 38928,
"runtimeRoot": "C:/Users/wangy/AppData/Local/Programs/AgentWorkbench/versions/Agent-Workbench-1.20.2-win-x64-20261101124548744",
"configSha256": "8a367d0eadafa696246471d616f8aaf34d2a72a9ebc1336803ad2da85f7b872b",
"projectsSha256": "bd6fe8fd6ff1d853e46e9c885be0b56a30cc146deaadd390a942124ab79043ae",
"sessionsSha256": "3142ff516faac454d7c7b1d13b219296afbcd0dc49f5273e46b2a0f7c3592a6e",
"modelCenterSha256": "f3f5839ab46b2b85f576bd7eaddd9160feb6509c76a2150eb7fa0c8d4c94b2d9",
"modelCenter": {
"revision": 3,
"selected": {
"type": "pool",
"id": "698ac145-cef9-409e-a09f-ed720143f7da"
},
"waitSettings": {
"modelResponseTimeoutMs": 300000,
"modelStreamIdleTimeoutMs": 120000
},
"connections": [
{
"id": "legacy-default",
"name": "原模型连接",
"brand": "兼容服务",
"protocol": "chat-completions",
"baseUrl": "http://127.0.0.1:11434/v1",
"model": "",
"enabled": true,
"requiresKey": false,
"revision": 2,
"contextLimit": 16000,
"maxOutputTokens": 2048,
"concurrency": 1,
"source": "legacy",
"hasKey": false,
"credentialRevision": 1,
"credentialStorage": "memory",
"status": "configured",
"activeRequests": 0,
"health": {
"state": "unverified"
},
"streaming": false,
"capabilitySchema": 1,
"capabilityDeclarations": {
"text": "supported",
"tools": "supported",
"streaming": "unknown",
"vision": "unknown",
"audio": "unknown"
},
"capabilities": {
"text": "protocol_declared",
"tools": "protocol_declared",
"streaming": "unknown",
"vision": "unknown",
"audio": "unknown"
}
},
{
"id": "62067dae-e266-47b6-b760-c6c48daa91f3",
"name": "Owned retained credential",
"brand": "自定义",
"protocol": "chat-completions",
"model": "owned-model",
"capabilitySchema": 1,
"capabilityDeclarations": {
"text": "supported",
"tools": "supported",
"streaming": "unknown",
"vision": "unknown",
"audio": "unknown"
},
"streaming": false,
"baseUrl": "http://127.0.0.1:1/v1",
"enabled": true,
"requiresKey": true,
"contextLimit": 16000,
"maxOutputTokens": 2048,
"concurrency": 1,
"revision": 1,
"source": "manual",
"createdAt": "2026-11-01T14:25:51.208Z",
"updatedAt": "2026-11-01T14:25:51.208Z",
"hasKey": true,
"credentialRevision": 1,
"credentialStorage": "memory",
"status": "configured",
"activeRequests": 0,
"health": {
"state": "unverified"
},
"capabilities": {
"text": "protocol_declared",
"tools": "protocol_declared",
"streaming": "unknown",
"vision": "unknown",
"audio": "unknown"
}
}
],
"pools": [
{
"id": "698ac145-cef9-409e-a09f-ed720143f7da",
"name": "Owned retained pool",
"members": [
{
"connectionId": "62067dae-e266-47b6-b760-c6c48daa91f3",
"weight": 1
}
],
"strategy": "round-robin",
"sticky": true,
"maxAttempts": 3,
"queueTimeoutMs": 15000,
"revision": 1
}
],
"imports": [],
"attempts": [],
"waiting": 0,
"storage": {
"credentials": "memory",
"restart": "临时凭据在宿主退出后需要重新填写;已明确本机加密保存的连接按用户与原配置加载。",
"attemptsRetained": 1000
}
},
"projects": 1,
"visibleSessions": 1,
"hasKey": false,
"active": [],
"storage": {
"engine": "sqlite",
"schemaVersion": 1,
"eventCursor": 1,
"pendingPublications": 1,
"migratedFromJson": false,
"recovery": "verify-unknown-results",
"compatibilityExport": "explicit-json"
},
"eventCursor": 1,
"database": {
"tables": {
"events": {
"count": 1,
"schemaSha256": "dd291c22596fd506ee7d7d588a3a55b81f01895930a069d1b4b30f42d77f2ce3",
"sha256": "3320f33a032518ef703adc57b5b780c288727436b750e93817af5cf97c69c916",
"rows": [
{
"id": "c9fd7b82-7abf-4f48-9d1d-05d85660faa8",
"sha256": "b57997798593b40945b9415834bef10ae5f3b5f2ebc8d6d10c6171159052c768"
}
]
},
"metadata": {
"count": 4,
"schemaSha256": "5a633f04efb47290eaac6bccc120405326eab37255b08b82685164f074400c8d",
"sha256": "ece152aca7cee9de5b892ea9e4405d897938be1f2057569499e8a28ceb1ce3b6",
"rows": [
{
"id": "config",
"sha256": "5b3cb0794e80dc1a6f247aec796a1a2e52161b1d791257aba623b3e88f7a1f1b"
},
{
"id": "migration",
"sha256": "f47011b196f85f6424a28141c6936e5d8463c1e80ebd49620b47f8757a3f7903"
},
{
"id": "initialized",
"sha256": "41e6ff8867f50debdbc0d60a83f90b181a552913f3787365d228270f00cecaa5"
},
{
"id": "modelRegistry",
"sha256": "ca7dfe2265a0a43d8d0d0097b2b6f5ac8d4ff073dfeeec8b8916796005c8e707"
}
]
},
"outbox": {
"count": 1,
"schemaSha256": "c645eeb96ba002abaf61e91dc503d967bffbd10a935b7416dd023386b9381552",
"sha256": "b9ee0ebd01fc20949b3878cf91d03016906d9c3940af970a02d08de1274d7b7c",
"rows": [
{
"id": 1,
"sha256": "eb5319436919e6d1789cb38514b45acf15713f0f3e919ac3be619bc9912dc84c"
}
]
},
"projects": {
"count": 1,
"schemaSha256": "c23c6b9539c0a20d9b52e7f7e98ddf7e532d7ee2ab85a911b904044ea8658737",
"sha256": "ec97e374c9229323ddda5cb0bbeb7eb35b1fb42acc494db57c64be07a5018879",
"rows": [
{
"id": "5b7848d9-d5ff-40d7-98d9-eb49c40c52c1",
"sha256": "1b9ea3fa0f26cbec79f2534155f1ea74294d8cc0699163d5619ee0fbbf1f8638"
}
]
},
"records": {
"count": 0,
"schemaSha256": "e64e4ef076c48464abf1973a8268461fb781ad99d0f0696c2f42d285c5369367",
"sha256": "4f53cda18c2baa0c0354bb5f9a3ecbe5ed12ab4d8e11ba873c2f11161202b945",
"rows": []
},
"sessions": {
"count": 1,
"schemaSha256": "f589250100bf72eb1440c43eb0439417b498de41882f3a3b650c721015c0a5ef",
"sha256": "f68476eb3a22458677c488833934ef252ceee0b40de64dc3adc3cc4fbccaee4e",
"rows": [
{
"id": "d29227d1-9ed9-407c-9116-53c54219728f",
"sha256": "44acdac4d2337c045108ea02e94d9232a84ee1492f1ca771402190b8c85f30c9"
}
]
},
"sqlite_sequence": {
"count": 1,
"schemaSha256": "4cb1eaf14467f226196148cb5688569660cb290d414bae4c1c450b149b62befd",
"sha256": "7ffc1e529c6a15c956cbbd1f9a069205567a7d65dd4ba629281aa405eb56d1ab",
"rows": [
{
"id": 1,
"sha256": "c0c433e6c0718fc96efc241368e9cbbb03487d5c427588f2efad69e64bad03f8"
}
]
}
},
"sessionIds": [
"d29227d1-9ed9-407c-9116-53c54219728f"
],
"ownerSha256": "f6fac6de050a5d6d4090bea7c1cf92ac40bca92b1dfff224ccc3d1e69d324c63",
"scope": "Every original table schema and row field hashed, excluding only runtimeOwner lifecycle metadata. No row payload or credential is returned."
},
"exports": [
{
"id": "d29227d1-9ed9-407c-9116-53c54219728f",
"project": "2ecb2d558741ae330d04d29e25bdb67c84f263ba6023c77dab608498f27a2500",
"session": "e52d591b846adf03ab9284c0d0010a988686ff2bbe0d11a8e0757741bc804e7a",
"execution": "e0f6d8b4dd63bf43d58681dbf19b5875da62545cc8913afa7dabaf5f45dac591"
}
],
"persistent": {
"tables": {
"events": {
"count": 1,
"hashes": [
"b57997798593b40945b9415834bef10ae5f3b5f2ebc8d6d10c6171159052c768"
]
},
"metadata": {
"count": 4,
"hashes": [
"41e6ff8867f50debdbc0d60a83f90b181a552913f3787365d228270f00cecaa5",
"5b3cb0794e80dc1a6f247aec796a1a2e52161b1d791257aba623b3e88f7a1f1b",
"ca7dfe2265a0a43d8d0d0097b2b6f5ac8d4ff073dfeeec8b8916796005c8e707",
"f47011b196f85f6424a28141c6936e5d8463c1e80ebd49620b47f8757a3f7903"
]
},
"outbox": {
"count": 1,
"hashes": [
"eb5319436919e6d1789cb38514b45acf15713f0f3e919ac3be619bc9912dc84c"
]
},
"projects": {
"count": 1,
"hashes": [
"1b9ea3fa0f26cbec79f2534155f1ea74294d8cc0699163d5619ee0fbbf1f8638"
]
},
"records": {
"count": 0,
"hashes": []
},
"sessions": {
"count": 1,
"hashes": [
"44acdac4d2337c045108ea02e94d9232a84ee1492f1ca771402190b8c85f30c9"
]
}
},
"artifacts": [],
"files": [
{
"path": "config.json",
"sha256": "6a7554aed121626c74eecd1c960c9378de8faf428b9bb3c8329d623adf6e0ff6"
}
],
"scope": "Every persistent database row except runtimeOwner, every original artifact byte and existing public config/project files; no credential is decrypted or returned."
},
"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."
},
"after": {
"observedAt": "2026-11-01T14:25:51.863Z",
"version": "1.20.3",
"pid": 38928,
"runtimeRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.20.3-win-x64-20261101141644779",
"configSha256": "8a367d0eadafa696246471d616f8aaf34d2a72a9ebc1336803ad2da85f7b872b",
"projectsSha256": "bd6fe8fd6ff1d853e46e9c885be0b56a30cc146deaadd390a942124ab79043ae",
"sessionsSha256": "3142ff516faac454d7c7b1d13b219296afbcd0dc49f5273e46b2a0f7c3592a6e",
"modelCenterSha256": "f3f5839ab46b2b85f576bd7eaddd9160feb6509c76a2150eb7fa0c8d4c94b2d9",
"modelCenter": {
"revision": 3,
"selected": {
"type": "pool",
"id": "698ac145-cef9-409e-a09f-ed720143f7da"
},
"waitSettings": {
"modelResponseTimeoutMs": 300000,
"modelStreamIdleTimeoutMs": 120000
},
"connections": [
{
"id": "legacy-default",
"name": "原模型连接",
"brand": "兼容服务",
"protocol": "chat-completions",
"baseUrl": "http://127.0.0.1:11434/v1",
"model": "",
"enabled": true,
"requiresKey": false,
"revision": 2,
"contextLimit": 16000,
"maxOutputTokens": 2048,
"concurrency": 1,
"source": "legacy",
"hasKey": false,
"credentialRevision": 1,
"credentialStorage": "memory",
"status": "configured",
"activeRequests": 0,
"health": {
"state": "unverified"
},
"streaming": false,
"capabilitySchema": 1,
"capabilityDeclarations": {
"text": "supported",
"tools": "supported",
"streaming": "unknown",
"vision": "unknown",
"audio": "unknown"
},
"capabilities": {
"text": "protocol_declared",
"tools": "protocol_declared",
"streaming": "unknown",
"vision": "unknown",
"audio": "unknown"
}
},
{
"id": "62067dae-e266-47b6-b760-c6c48daa91f3",
"name": "Owned retained credential",
"brand": "自定义",
"protocol": "chat-completions",
"model": "owned-model",
"capabilitySchema": 1,
"capabilityDeclarations": {
"text": "supported",
"tools": "supported",
"streaming": "unknown",
"vision": "unknown",
"audio": "unknown"
},
"streaming": false,
"baseUrl": "http://127.0.0.1:1/v1",
"enabled": true,
"requiresKey": true,
"contextLimit": 16000,
"maxOutputTokens": 2048,
"concurrency": 1,
"revision": 1,
"source": "manual",
"createdAt": "2026-11-01T14:25:51.208Z",
"updatedAt": "2026-11-01T14:25:51.208Z",
"hasKey": true,
"credentialRevision": 1,
"credentialStorage": "memory",
"status": "configured",
"activeRequests": 0,
"health": {
"state": "unverified"
},
"capabilities": {
"text": "protocol_declared",
"tools": "protocol_declared",
"streaming": "unknown",
"vision": "unknown",
"audio": "unknown"
}
}
],
"pools": [
{
"id": "698ac145-cef9-409e-a09f-ed720143f7da",
"name": "Owned retained pool",
"members": [
{
"connectionId": "62067dae-e266-47b6-b760-c6c48daa91f3",
"weight": 1
}
],
"strategy": "round-robin",
"sticky": true,
"maxAttempts": 3,
"queueTimeoutMs": 15000,
"revision": 1
}
],
"imports": [],
"attempts": [],
"waiting": 0,
"storage": {
"credentials": "memory",
"restart": "临时凭据在宿主退出后需要重新填写;已明确本机加密保存的连接按用户与原配置加载。",
"attemptsRetained": 1000
}
},
"projects": 1,
"visibleSessions": 1,
"hasKey": false,
"active": [],
"storage": {
"engine": "sqlite",
"schemaVersion": 1,
"eventCursor": 1,
"pendingPublications": 1,
"migratedFromJson": false,
"recovery": "verify-unknown-results",
"compatibilityExport": "explicit-json"
},
"eventCursor": 1,
"database": {
"tables": {
"events": {
"count": 1,
"schemaSha256": "dd291c22596fd506ee7d7d588a3a55b81f01895930a069d1b4b30f42d77f2ce3",
"sha256": "3320f33a032518ef703adc57b5b780c288727436b750e93817af5cf97c69c916",
"rows": [
{
"id": "c9fd7b82-7abf-4f48-9d1d-05d85660faa8",
"sha256": "b57997798593b40945b9415834bef10ae5f3b5f2ebc8d6d10c6171159052c768"
}
]
},
"metadata": {
"count": 4,
"schemaSha256": "5a633f04efb47290eaac6bccc120405326eab37255b08b82685164f074400c8d",
"sha256": "ece152aca7cee9de5b892ea9e4405d897938be1f2057569499e8a28ceb1ce3b6",
"rows": [
{
"id": "config",
"sha256": "5b3cb0794e80dc1a6f247aec796a1a2e52161b1d791257aba623b3e88f7a1f1b"
},
{
"id": "migration",
"sha256": "f47011b196f85f6424a28141c6936e5d8463c1e80ebd49620b47f8757a3f7903"
},
{
"id": "initialized",
"sha256": "41e6ff8867f50debdbc0d60a83f90b181a552913f3787365d228270f00cecaa5"
},
{
"id": "modelRegistry",
"sha256": "ca7dfe2265a0a43d8d0d0097b2b6f5ac8d4ff073dfeeec8b8916796005c8e707"
}
]
},
"outbox": {
"count": 1,
"schemaSha256": "c645eeb96ba002abaf61e91dc503d967bffbd10a935b7416dd023386b9381552",
"sha256": "b9ee0ebd01fc20949b3878cf91d03016906d9c3940af970a02d08de1274d7b7c",
"rows": [
{
"id": 1,
"sha256": "eb5319436919e6d1789cb38514b45acf15713f0f3e919ac3be619bc9912dc84c"
}
]
},
"projects": {
"count": 1,
"schemaSha256": "c23c6b9539c0a20d9b52e7f7e98ddf7e532d7ee2ab85a911b904044ea8658737",
"sha256": "ec97e374c9229323ddda5cb0bbeb7eb35b1fb42acc494db57c64be07a5018879",
"rows": [
{
"id": "5b7848d9-d5ff-40d7-98d9-eb49c40c52c1",
"sha256": "1b9ea3fa0f26cbec79f2534155f1ea74294d8cc0699163d5619ee0fbbf1f8638"
}
]
},
"records": {
"count": 0,
"schemaSha256": "e64e4ef076c48464abf1973a8268461fb781ad99d0f0696c2f42d285c5369367",
"sha256": "4f53cda18c2baa0c0354bb5f9a3ecbe5ed12ab4d8e11ba873c2f11161202b945",
"rows": []
},
"sessions": {
"count": 1,
"schemaSha256": "f589250100bf72eb1440c43eb0439417b498de41882f3a3b650c721015c0a5ef",
"sha256": "f68476eb3a22458677c488833934ef252ceee0b40de64dc3adc3cc4fbccaee4e",
"rows": [
{
"id": "d29227d1-9ed9-407c-9116-53c54219728f",
"sha256": "44acdac4d2337c045108ea02e94d9232a84ee1492f1ca771402190b8c85f30c9"
}
]
},
"sqlite_sequence": {
"count": 1,
"schemaSha256": "4cb1eaf14467f226196148cb5688569660cb290d414bae4c1c450b149b62befd",
"sha256": "7ffc1e529c6a15c956cbbd1f9a069205567a7d65dd4ba629281aa405eb56d1ab",
"rows": [
{
"id": 1,
"sha256": "c0c433e6c0718fc96efc241368e9cbbb03487d5c427588f2efad69e64bad03f8"
}
]
}
},
"sessionIds": [
"d29227d1-9ed9-407c-9116-53c54219728f"
],
"ownerSha256": "f6fac6de050a5d6d4090bea7c1cf92ac40bca92b1dfff224ccc3d1e69d324c63",
"scope": "Every original table schema and row field hashed, excluding only runtimeOwner lifecycle metadata. No row payload or credential is returned."
},
"exports": [
{
"id": "d29227d1-9ed9-407c-9116-53c54219728f",
"project": "2ecb2d558741ae330d04d29e25bdb67c84f263ba6023c77dab608498f27a2500",
"session": "e52d591b846adf03ab9284c0d0010a988686ff2bbe0d11a8e0757741bc804e7a",
"execution": "e0f6d8b4dd63bf43d58681dbf19b5875da62545cc8913afa7dabaf5f45dac591"
}
],
"persistent": {
"tables": {
"events": {
"count": 1,
"hashes": [
"b57997798593b40945b9415834bef10ae5f3b5f2ebc8d6d10c6171159052c768"
]
},
"metadata": {
"count": 4,
"hashes": [
"41e6ff8867f50debdbc0d60a83f90b181a552913f3787365d228270f00cecaa5",
"5b3cb0794e80dc1a6f247aec796a1a2e52161b1d791257aba623b3e88f7a1f1b",
"ca7dfe2265a0a43d8d0d0097b2b6f5ac8d4ff073dfeeec8b8916796005c8e707",
"f47011b196f85f6424a28141c6936e5d8463c1e80ebd49620b47f8757a3f7903"
]
},
"outbox": {
"count": 1,
"hashes": [
"eb5319436919e6d1789cb38514b45acf15713f0f3e919ac3be619bc9912dc84c"
]
},
"projects": {
"count": 1,
"hashes": [
"1b9ea3fa0f26cbec79f2534155f1ea74294d8cc0699163d5619ee0fbbf1f8638"
]
},
"records": {
"count": 0,
"hashes": []
},
"sessions": {
"count": 1,
"hashes": [
"44acdac4d2337c045108ea02e94d9232a84ee1492f1ca771402190b8c85f30c9"
]
}
},
"artifacts": [],
"files": [
{
"path": "config.json",
"sha256": "6a7554aed121626c74eecd1c960c9378de8faf428b9bb3c8329d623adf6e0ff6"
}
],
"scope": "Every persistent database row except runtimeOwner, every original artifact byte and existing public config/project files; no credential is decrypted or returned."
},
"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."
},
"realModelRequests": 0,
"scope": "Actual old compiled host to new immutable compiled handler; own credential and pool, original PID/public config/every table/export/artifact retained. No request to model endpoint."
}

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