docs: bind final 1.14 release to source, portable and unchanged product evidence
This commit is contained in:
@@ -1,6 +1,8 @@
|
||||
# 本地 Agent 工作台
|
||||
|
||||
Agent Workbench 将 Claude Code 与 Cursor 的机制研究整合为独立的本地 AI 编程工作台,并内置模型池、项目记忆、Skills/MCP、REA 工件研究、本地DeepSeek Harness连接核验、实际页面验收和有依赖的执行计划。工程索引、精确补丁、多文件整批审查、Node/Python实际检查、浏览器交互、本地Git提交及交付报告可以在同一个Task中完成。
|
||||
Agent Workbench 将 Claude Code 与 Cursor 的机制研究整合为独立的本地 AI 编程工作台,并内置模型池、项目记忆、Skills/MCP、REA 工件研究、受控DeepSeek Harness工程引擎、实际页面验收和有依赖的执行计划。工程索引、精确补丁、多文件整批审查、Node/Python实际检查、浏览器交互、本地Git提交及交付报告可以在同一个Task中完成。
|
||||
|
||||
1.14.0新增固定版本的受控引擎与工具/模型桥,原生prompt、审批、预算、原结果和停止进入所属Task/Run;独立工作区成果由父工程审查、检查及Git交付。Purge两版本来源锁定的提示实验默认关闭。三个真实功能工程13节点、23类178组UI和独立源码445/445通过,随包Node/npm/REA/SDK/Edge与受控工程也已实测;最终发行及当前[本地工作台](http://127.0.0.1:4318/?release=1.14.0)状态见[QA-1.14](docs/QA-1.14.md)。真实模型推理收益按原选择独立待验。
|
||||
|
||||
1.13.0已本地交付[隔离协作](docs/WORKERS.md):父目标与两个受限Git工作区、累计预算/停止恢复、当前成果核对、具体冲突和父工程审查/检查/Git交付。独立源码404/404、完整界面21类160组与三个真实功能工程通过,父Node8/8、React9/9、Python10/10及构建/编译/Git,React内置Edge两视口各10项断言。正式[本地工作台](http://127.0.0.1:4318/?release=1.13.0)已保持原进程/模型连接/全部历史升级,验收与部署见[QA-1.13](docs/QA-1.13.md)。
|
||||
|
||||
|
||||
+1
-1
@@ -1,6 +1,6 @@
|
||||
# 1.14 受控外部引擎与增强策略执行方案
|
||||
|
||||
2026-10-31。此文记录1.13隔离协作之后的执行方案及接续状态。A–E受控引擎、工具桥、实际工程、恢复界面和默认关闭提示实验已实现并各自验收;F已补升级活动锁/回退、原宿主身份幂等和旧导出兼容,编译宿主完整工程通过,最终完整回归、独立发行与正式升级仍在推进。正式服务目前仍1.13.0;各轮原结果以[1.14验收](QA-1.14.md)与[回顾](RETROSPECTIVE-1.14.md)为准。
|
||||
2026-10-31。此文记录1.13隔离协作之后的执行方案及接续状态。A–E受控引擎、工具桥、实际工程、恢复界面和默认关闭提示实验已实现并各自验收;F升级活动锁/回退、原宿主身份幂等和旧导出兼容、三工程13节点、23类178组UI、独立445/445与随包完整工程已通过。两份QA脚本的路径/依赖修复和原失败分别封存,最终不可变发行及保留状态正式升级仍在进行。正式服务目前仍1.13.0;各轮原结果以[1.14验收](QA-1.14.md)与[回顾](RETROSPECTIVE-1.14.md)为准。
|
||||
|
||||
## 要交付的能力
|
||||
|
||||
|
||||
@@ -1,8 +1,8 @@
|
||||
{
|
||||
"success": true,
|
||||
"version": "1.14.0",
|
||||
"verifiedAt": "2026-10-30T23:20:30.732Z",
|
||||
"codeCommit": "57d075ef5b09f70cf3a7b296df5d48c4b740b233",
|
||||
"verifiedAt": "2026-10-30T23:53:04.437Z",
|
||||
"codeCommit": "b3111160e24c78325a0888b112fb334e30778756",
|
||||
"frozenInputs": 253,
|
||||
"builds": 6,
|
||||
"benchmarks": 40,
|
||||
@@ -104,5 +104,43 @@
|
||||
"scope": "Only the compiled acceptance verifier allocates its synthetic evidence/Git project in an exclusive OS temporary directory. Every product source, test, UI verifier, benchmark and built byte remains equal to the original 53e196a complete project/UI acceptance; no workspace guard or completion assertion changed.",
|
||||
"realModelRequests": 0
|
||||
},
|
||||
"postSourceValidationOnlyChange": {
|
||||
"success": true,
|
||||
"version": "1.14.0",
|
||||
"recordedAt": "2026-10-30T23:51:55.456Z",
|
||||
"commit": "b3111160e24c78325a0888b112fb334e30778756",
|
||||
"productCodeCommit": "53e196af5954853bab37f03b7ca34fd55a902b3e",
|
||||
"sourceValidatedCommit": "57d075ef5b09f70cf3a7b296df5d48c4b740b233",
|
||||
"sourceTests": {
|
||||
"tests": 445,
|
||||
"pass": 445,
|
||||
"fail": 0,
|
||||
"cancelled": 0,
|
||||
"skipped": 0,
|
||||
"todo": 0
|
||||
},
|
||||
"change": {
|
||||
"path": "scripts/verify-release.mjs",
|
||||
"originalSha256": "5147a52e8fbe0cc5f10825479a3b1e75c208ae671ec7a21d5bc97849dbb8be46",
|
||||
"sha256": "987dca1eae2077916f3a62ec58f6b59a4faeec4092b53a5c1e376ff25b8c8ab1",
|
||||
"originalBytes": 15164,
|
||||
"bytes": 16062
|
||||
},
|
||||
"unchangedTestedInputs": 252,
|
||||
"unchangedBuilds": 6,
|
||||
"unchangedBenchmarks": 40,
|
||||
"externalGit": {
|
||||
"path": "C:\\Program Files\\Git\\cmd\\git.exe",
|
||||
"version": "git version 2.54.0.windows.1",
|
||||
"sha256": "81ef35ae005ca9318018d18e3327578ce939fb99feaad6b2d7c8ab15f3de8db5"
|
||||
},
|
||||
"actualPortableRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.14.0-win-x64-20261030232056161",
|
||||
"portableEngineCases": 4,
|
||||
"portableFixtureRequests": 7,
|
||||
"portableProofPath": "docs/controlled-engine-portable-git-fix-validation.json",
|
||||
"portableProofSha256": "06d0f141317c451c5481b6370c98ca227e86940615a17a0b1b67c5974fef8636",
|
||||
"scope": "Only the portable verifier PATH adds the exact observed external Git directory; no system Node/npm is available. The completed 445-test source run remains tied to its original inputs. This one post-source QA script change is exercised by actual portable Node/npm/REA/SDK/Edge/native repair/review/Git; every product/test/UI/benchmark/built byte is unchanged.",
|
||||
"realModelRequests": 0
|
||||
},
|
||||
"scope": "Rechecks existing original bytes and actual raw engine/native cleanup/browser screenshot contracts. It does not rerun projects or UI tests."
|
||||
}
|
||||
|
||||
@@ -7,8 +7,10 @@ import { execFileSync } from 'node:child_process';
|
||||
const read = filename => JSON.parse(fs.readFileSync(filename, 'utf8'));
|
||||
const hash = bytes => createHash('sha256').update(bytes).digest('hex');
|
||||
const originalBaseline = read('docs/controlled-engine-code-baseline.json');
|
||||
const baseline = fs.existsSync('docs/controlled-engine-final-code-baseline.json') ? read('docs/controlled-engine-final-code-baseline.json') : originalBaseline;
|
||||
const baseline = fs.existsSync('docs/controlled-engine-final-release-baseline.json') ? read('docs/controlled-engine-final-release-baseline.json')
|
||||
: fs.existsSync('docs/controlled-engine-final-code-baseline.json') ? read('docs/controlled-engine-final-code-baseline.json') : originalBaseline;
|
||||
const qaChange = fs.existsSync('docs/controlled-engine-qa-path-fix-validation.json') ? read('docs/controlled-engine-qa-path-fix-validation.json') : null;
|
||||
const portableQa = fs.existsSync('docs/controlled-engine-portable-qa-change.json') ? read('docs/controlled-engine-portable-qa-change.json') : null;
|
||||
const projects = read('docs/controlled-engine-final-projects-validation.json');
|
||||
const ui = read('docs/controlled-engine-all-ui-validation.json');
|
||||
const logs = read('docs/controlled-engine-final-ui-logs.json');
|
||||
@@ -26,6 +28,10 @@ for (const row of projects.inputs) {
|
||||
assert.equal(qaChange.success, true); assert.equal(qaChange.change.path, 'scripts/verify-controlled-release.mjs');
|
||||
assert.equal(row.sha256, qaChange.change.originalSha256); assert.equal(hash(fs.readFileSync(row.path)), qaChange.change.sha256);
|
||||
assert.equal(hash(execFileSync('git', ['show', originalBaseline.codeCommit + ':' + filename], { windowsHide: true, maxBuffer: 1048576 })), row.sha256);
|
||||
} else if (portableQa && filename === portableQa.change.path) {
|
||||
assert.equal(portableQa.success, true); assert.equal(portableQa.change.path, 'scripts/verify-release.mjs');
|
||||
assert.equal(row.sha256, portableQa.change.originalSha256); assert.equal(hash(fs.readFileSync(row.path)), portableQa.change.sha256);
|
||||
assert.equal(hash(execFileSync('git', ['show', portableQa.sourceValidatedCommit + ':' + filename], { windowsHide: true, maxBuffer: 1048576 })), row.sha256);
|
||||
} else assert.equal(hash(fs.readFileSync(row.path)), row.sha256, row.path);
|
||||
}
|
||||
for (const row of logs.logs) {
|
||||
@@ -59,7 +65,7 @@ for (const [index, sample] of projects.results.entries()) {
|
||||
const result = { success: true, version: '1.14.0', verifiedAt: new Date().toISOString(), codeCommit: baseline.codeCommit,
|
||||
frozenInputs: baseline.inputs.length, builds: baseline.builds.length, benchmarks: baseline.benchmarks.length,
|
||||
projectInputFiles: projects.inputs.length, projectCases: projects.cases.length, projects: summaries,
|
||||
uiTypes: ui.results.length, uiGroups: ui.groups, exactUiLogs: logs.logs.length, images, realModelRequests: 0, validationOnlyChange: qaChange,
|
||||
uiTypes: ui.results.length, uiGroups: ui.groups, exactUiLogs: logs.logs.length, images, realModelRequests: 0, validationOnlyChange: qaChange, postSourceValidationOnlyChange: portableQa,
|
||||
scope: 'Rechecks existing original bytes and actual raw engine/native cleanup/browser screenshot contracts. It does not rerun projects or UI tests.' };
|
||||
fs.writeFileSync('docs/controlled-engine-evidence-validation.json', JSON.stringify(result, null, 2) + '\n');
|
||||
console.log(JSON.stringify({ success: true, version: result.version, uiGroups: result.uiGroups, projectCases: result.projectCases,
|
||||
|
||||
File diff suppressed because it is too large
Load Diff
@@ -0,0 +1,38 @@
|
||||
{
|
||||
"success": true,
|
||||
"version": "1.14.0",
|
||||
"recordedAt": "2026-10-30T23:51:55.456Z",
|
||||
"commit": "b3111160e24c78325a0888b112fb334e30778756",
|
||||
"productCodeCommit": "53e196af5954853bab37f03b7ca34fd55a902b3e",
|
||||
"sourceValidatedCommit": "57d075ef5b09f70cf3a7b296df5d48c4b740b233",
|
||||
"sourceTests": {
|
||||
"tests": 445,
|
||||
"pass": 445,
|
||||
"fail": 0,
|
||||
"cancelled": 0,
|
||||
"skipped": 0,
|
||||
"todo": 0
|
||||
},
|
||||
"change": {
|
||||
"path": "scripts/verify-release.mjs",
|
||||
"originalSha256": "5147a52e8fbe0cc5f10825479a3b1e75c208ae671ec7a21d5bc97849dbb8be46",
|
||||
"sha256": "987dca1eae2077916f3a62ec58f6b59a4faeec4092b53a5c1e376ff25b8c8ab1",
|
||||
"originalBytes": 15164,
|
||||
"bytes": 16062
|
||||
},
|
||||
"unchangedTestedInputs": 252,
|
||||
"unchangedBuilds": 6,
|
||||
"unchangedBenchmarks": 40,
|
||||
"externalGit": {
|
||||
"path": "C:\\Program Files\\Git\\cmd\\git.exe",
|
||||
"version": "git version 2.54.0.windows.1",
|
||||
"sha256": "81ef35ae005ca9318018d18e3327578ce939fb99feaad6b2d7c8ab15f3de8db5"
|
||||
},
|
||||
"actualPortableRoot": "D:\\Ai-project\\REA逆向\\Agent-Workbench\\releases\\Agent-Workbench-1.14.0-win-x64-20261030232056161",
|
||||
"portableEngineCases": 4,
|
||||
"portableFixtureRequests": 7,
|
||||
"portableProofPath": "docs/controlled-engine-portable-git-fix-validation.json",
|
||||
"portableProofSha256": "06d0f141317c451c5481b6370c98ca227e86940615a17a0b1b67c5974fef8636",
|
||||
"scope": "Only the portable verifier PATH adds the exact observed external Git directory; no system Node/npm is available. The completed 445-test source run remains tied to its original inputs. This one post-source QA script change is exercised by actual portable Node/npm/REA/SDK/Edge/native repair/review/Git; every product/test/UI/benchmark/built byte is unchanged.",
|
||||
"realModelRequests": 0
|
||||
}
|
||||
@@ -586,3 +586,6 @@ Claude Code 报告已验证工具输入校验、权限决定、队列并发屏
|
||||
|
||||
- 首个独立源码候选在受控编译工程阶段失败:验证工程嵌套于长解压目录,原Windows Git路径守卫拒绝(stale),完整源码测试尚未执行。六阶段原日志、原失败/候选/归档以及已正常exit0的所属宿主已原字节保存docs/controlled-engine-source-*,不计候选通过。
|
||||
- 仅修正verify-controlled-release验收目录为exclusive OS temporary目录,与安装深度分开;所有产品源码、检查断言、路径保护、旧完整UI/三工程和构建字节保持。先在原长安装目录验证完整编译链,再重新冻结QA输入/独立发行。新增baseline/差异账本明确这一份验证脚本变化,不覆盖原53e196a历史输入。
|
||||
|
||||
- 独立完整源码原终态445/445通过,0失败/取消/跳过/todo;7阶段原日志与253输入/6构建/8,836份源码及依赖比对封存,249输入Git原字节相同、4份旧换行差异保留。会话36928已正常结束,不再重跑整轮。
|
||||
- 便携PATH缺Git的真实ENOENT复现已保存;b311116只修验收脚本显式加入Git2.54.0已观测目录。首个完整便携/随包Node/npm4项、REA/SDK/Edge及受控四节点、7次本地HTTP/原8项检查/父审查/Git与完整停止通过。新final-release-baseline和portable-qa-change明确后于源码回归的一份QA文件变化,source原输入不改写,所有产品/测试/UI/基准/构建保持。准备最终不可变包与保留状态部署。
|
||||
|
||||
@@ -610,3 +610,6 @@
|
||||
- 仅修正verify-controlled-release验收目录为exclusive OS temporary目录,与安装深度分开;所有产品源码、检查断言、路径保护、旧完整UI/三工程和构建字节保持。先在原长安装目录验证完整编译链,再重新冻结QA输入/独立发行。新增baseline/差异账本明确这一份验证脚本变化,不覆盖原53e196a历史输入。
|
||||
|
||||
- 1.14新独立源码完整回归会话36928仍运行,截至375项均通过,未结束不计整轮成功。离线安装/构建、REA、SDK、Edge及受控编译四节点已完成;原源码目录/7阶段日志保持,最终统计以原TAP尾部和source-package-validation终态为准。正式1.13/PID23172保持。
|
||||
|
||||
- 独立完整源码原终态445/445通过,0失败/取消/跳过/todo;7阶段原日志与253输入/6构建/8,836份源码及依赖比对封存,249输入Git原字节相同、4份旧换行差异保留。会话36928已正常结束,不再重跑整轮。
|
||||
- 便携PATH缺Git的真实ENOENT复现已保存;b311116只修验收脚本显式加入Git2.54.0已观测目录。首个完整便携/随包Node/npm4项、REA/SDK/Edge及受控四节点、7次本地HTTP/原8项检查/父审查/Git与完整停止通过。新final-release-baseline和portable-qa-change明确后于源码回归的一份QA文件变化,source原输入不改写,所有产品/测试/UI/基准/构建保持。准备最终不可变包与保留状态部署。
|
||||
|
||||
+1
-1
@@ -13,7 +13,7 @@
|
||||
|
||||
初步核对:原SDK没有session/cancel,prompt只给入队收据,client/server反向请求审批未实现;sdk-minimal仍包含本机文件/终端工具及danger-full-access策略,不能直接赋予工程写入。本阶段通过自有版本化叠加插件与受限宿主桥接入,实际工具目录和所属账本是验收门槛。完整公开源码使用普通代码工具;需要发行资源与运行观察时复用REA。
|
||||
|
||||
当前F接续:57d075e只修正verify-controlled-release的合成Git工程目录,长安装目录四节点复验通过,6d09ac8封存原/新SHA及所有未变化产品/UI/基准/构建。新不可变20261030232056161独立源码会话36928运行中,先核对原结果,不重启已有测试。随后完成文档封存、便携及保留正式23172/Provider/所有行与导出的升级。
|
||||
当前F接续:57d075e修正独立编译QA路径,6d09ac8封存差异;20261030232056161独立445/445及7阶段全部通过,253输入/6构建/8,836份运行依赖已核对。b311116修正便携QA误移除外部Git,实际完整便携/受控四节点通过;仅这一份源码整轮之后的QA脚本变化另行封存,产品/测试/UI/基准/构建保持。现生成最终不可变发行,再保留正式23172/Provider/所有行与导出升级及14页复验。
|
||||
|
||||
本轮QA错误:nativeSummary被误用于raw engine_run、project输入快照的未记录bytes被当必有字段、首个编译验收工程路径过深;分别按实际原schema与exclusive临时工程修正,原失败/日志保留。一次只读verify-archives.mjs和worker-fixture路径猜测不存在,已按实际verify-archives.py及现有导入核对。没有放宽产品权限、路径、完成或进程关闭断言。
|
||||
|
||||
|
||||
Reference in New Issue
Block a user