coding-agent-harness 1.0.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +13 -0
- package/LICENSE +21 -0
- package/README.md +141 -0
- package/SKILL.md +423 -0
- package/docs-release/README.md +30 -0
- package/docs-release/architecture/overview.md +52 -0
- package/docs-release/guides/agent-installation.md +139 -0
- package/examples/minimal-project/.harness-capabilities.json +8 -0
- package/examples/minimal-project/AGENTS.md +4 -0
- package/examples/minimal-project/CLAUDE.md +3 -0
- package/examples/minimal-project/docs/09-PLANNING/TASKS/demo-task/execution_strategy.md +10 -0
- package/examples/minimal-project/docs/09-PLANNING/TASKS/demo-task/progress.md +11 -0
- package/examples/minimal-project/docs/09-PLANNING/TASKS/demo-task/review.md +27 -0
- package/examples/minimal-project/docs/09-PLANNING/TASKS/demo-task/task_plan.md +14 -0
- package/examples/minimal-project/docs/09-PLANNING/TASKS/demo-task/visual_roadmap.md +11 -0
- package/examples/minimal-project/docs/Harness-Ledger.md +6 -0
- package/package.json +34 -0
- package/references/adversarial-review-standard.md +173 -0
- package/references/agents-md-pattern.md +140 -0
- package/references/cadence-ledger.md +55 -0
- package/references/ci-cd-standard.md +90 -0
- package/references/delivery-operating-model-standard.md +145 -0
- package/references/docs-directory-standard.md +125 -0
- package/references/harness-ledger.md +148 -0
- package/references/lessons-governance.md +157 -0
- package/references/long-running-task-standard.md +209 -0
- package/references/module-parallel-standard.md +292 -0
- package/references/planning-loop.md +192 -0
- package/references/project-onboarding-audit.md +167 -0
- package/references/regression-system.md +89 -0
- package/references/repo-governance-standard.md +131 -0
- package/references/review-routing-standard.md +103 -0
- package/references/ssot-governance.md +111 -0
- package/references/walkthrough-closeout.md +135 -0
- package/references/worktree-parallel.md +184 -0
- package/scripts/check-harness.mjs +728 -0
- package/scripts/harness.mjs +201 -0
- package/scripts/lib/dashboard-writer.mjs +95 -0
- package/scripts/lib/harness-core.mjs +1318 -0
- package/scripts/smoke-dashboard.mjs +70 -0
- package/scripts/test-harness.mjs +482 -0
- package/templates/AGENTS.md.template +82 -0
- package/templates/CLAUDE.md.template +12 -0
- package/templates/dashboard/assets/app.css +399 -0
- package/templates/dashboard/assets/app.js +435 -0
- package/templates/dashboard/assets/i18n.js +47 -0
- package/templates/dashboard/assets/markdown-reader.js +116 -0
- package/templates/dashboard/assets/mermaid-renderer.js +59 -0
- package/templates/dashboard/index.html +18 -0
- package/templates/ledger/Harness-Ledger.md +39 -0
- package/templates/lessons/lesson-arch-process-change.md +47 -0
- package/templates/lessons/lesson-new-doc.md +50 -0
- package/templates/lessons/lesson-ref-change.md +45 -0
- package/templates/planning/execution_strategy.md +40 -0
- package/templates/planning/findings.md +24 -0
- package/templates/planning/long-running-task-contract.md +69 -0
- package/templates/planning/module_plan.md +36 -0
- package/templates/planning/module_session_prompt.md +39 -0
- package/templates/planning/optional/artifacts/INDEX.md +12 -0
- package/templates/planning/optional/references/INDEX.md +13 -0
- package/templates/planning/optional/slices/_slice-template/brief.md +27 -0
- package/templates/planning/optional/slices/_slice-template/evidence.md +9 -0
- package/templates/planning/optional/slices/_slice-template/review.md +31 -0
- package/templates/planning/progress.md +33 -0
- package/templates/planning/review.md +48 -0
- package/templates/planning/task_plan.md +86 -0
- package/templates/planning/visual_roadmap.md +28 -0
- package/templates/reference/adversarial-review-standard.md +28 -0
- package/templates/reference/ci-cd-standard.md +28 -0
- package/templates/reference/delivery-operating-model-standard.md +28 -0
- package/templates/reference/docs-library-standard.md +28 -0
- package/templates/reference/engineering-standard.md +29 -0
- package/templates/reference/execution-workflow-standard.md +29 -0
- package/templates/reference/harness-ledger-standard.md +26 -0
- package/templates/reference/long-running-task-standard.md +28 -0
- package/templates/reference/regression-ssot-governance.md +28 -0
- package/templates/reference/repo-governance-standard.md +29 -0
- package/templates/reference/review-routing-standard.md +29 -0
- package/templates/reference/testing-standard.md +28 -0
- package/templates/reference/walkthrough-standard.md +28 -0
- package/templates/reference/worktree-standard.md +28 -0
- package/templates/regression/Cadence-Ledger.md +41 -0
- package/templates/ssot/Delivery-SSoT.md +43 -0
- package/templates/ssot/Feature-SSoT.md +43 -0
- package/templates/ssot/Lessons-SSoT.md +44 -0
- package/templates/ssot/Module-Registry.md +43 -0
- package/templates/ssot/Regression-SSoT.md +51 -0
- package/templates/verifier/verifier-output.md +43 -0
- package/templates/walkthrough/Closeout-SSoT.md +43 -0
- package/templates/walkthrough/walkthrough-template.md +63 -0
- package/templates-zh-CN/AGENTS.md.template +92 -0
- package/templates-zh-CN/CLAUDE.md.template +12 -0
- package/templates-zh-CN/dashboard/assets/app.css +399 -0
- package/templates-zh-CN/dashboard/assets/app.js +435 -0
- package/templates-zh-CN/dashboard/assets/i18n.js +47 -0
- package/templates-zh-CN/dashboard/assets/markdown-reader.js +116 -0
- package/templates-zh-CN/dashboard/assets/mermaid-renderer.js +59 -0
- package/templates-zh-CN/dashboard/index.html +18 -0
- package/templates-zh-CN/ledger/Harness-Ledger.md +50 -0
- package/templates-zh-CN/lessons/lesson-arch-process-change.md +47 -0
- package/templates-zh-CN/lessons/lesson-new-doc.md +49 -0
- package/templates-zh-CN/lessons/lesson-ref-change.md +59 -0
- package/templates-zh-CN/planning/execution_strategy.md +37 -0
- package/templates-zh-CN/planning/findings.md +24 -0
- package/templates-zh-CN/planning/long-running-task-contract.md +118 -0
- package/templates-zh-CN/planning/module_plan.md +43 -0
- package/templates-zh-CN/planning/module_session_prompt.md +70 -0
- package/templates-zh-CN/planning/optional/artifacts/INDEX.md +13 -0
- package/templates-zh-CN/planning/optional/references/INDEX.md +13 -0
- package/templates-zh-CN/planning/optional/slices/_slice-template/brief.md +35 -0
- package/templates-zh-CN/planning/optional/slices/_slice-template/evidence.md +12 -0
- package/templates-zh-CN/planning/optional/slices/_slice-template/review.md +37 -0
- package/templates-zh-CN/planning/progress.md +29 -0
- package/templates-zh-CN/planning/review.md +69 -0
- package/templates-zh-CN/planning/task_plan.md +116 -0
- package/templates-zh-CN/planning/visual_roadmap.md +24 -0
- package/templates-zh-CN/reference/adversarial-review-standard.md +89 -0
- package/templates-zh-CN/reference/ci-cd-standard.md +72 -0
- package/templates-zh-CN/reference/delivery-operating-model-standard.md +79 -0
- package/templates-zh-CN/reference/docs-library-standard.md +59 -0
- package/templates-zh-CN/reference/engineering-standard.md +80 -0
- package/templates-zh-CN/reference/execution-workflow-standard.md +81 -0
- package/templates-zh-CN/reference/harness-ledger-standard.md +91 -0
- package/templates-zh-CN/reference/long-running-task-standard.md +156 -0
- package/templates-zh-CN/reference/regression-ssot-governance.md +82 -0
- package/templates-zh-CN/reference/repo-governance-standard.md +84 -0
- package/templates-zh-CN/reference/review-routing-standard.md +82 -0
- package/templates-zh-CN/reference/testing-standard.md +72 -0
- package/templates-zh-CN/reference/walkthrough-standard.md +83 -0
- package/templates-zh-CN/reference/worktree-standard.md +116 -0
- package/templates-zh-CN/regression/Cadence-Ledger.md +48 -0
- package/templates-zh-CN/ssot/Delivery-SSoT.md +60 -0
- package/templates-zh-CN/ssot/Feature-SSoT.md +49 -0
- package/templates-zh-CN/ssot/Lessons-SSoT.md +49 -0
- package/templates-zh-CN/ssot/Module-Registry.md +48 -0
- package/templates-zh-CN/ssot/Regression-SSoT.md +51 -0
- package/templates-zh-CN/verifier/verifier-output.md +38 -0
- package/templates-zh-CN/walkthrough/Closeout-SSoT.md +42 -0
- package/templates-zh-CN/walkthrough/walkthrough-template.md +62 -0
|
@@ -0,0 +1,192 @@
|
|
|
1
|
+
# Planning Loop
|
|
2
|
+
|
|
3
|
+
## 核心思路
|
|
4
|
+
|
|
5
|
+
每个非平凡任务必须有独立的任务目录和稳定的任务文件。这是 agent 在长时间执行过程中不偏离目标的锚点。
|
|
6
|
+
|
|
7
|
+
## 任务目录结构
|
|
8
|
+
|
|
9
|
+
```
|
|
10
|
+
docs/09-PLANNING/TASKS/<YYYY-MM-DD-任务名>/
|
|
11
|
+
├── task_plan.md ← 计划:目标、范围、步骤、验收标准
|
|
12
|
+
├── execution_strategy.md ← 执行策略:模式、subagent、冲突控制、验证深度、handoff
|
|
13
|
+
├── visual_roadmap.md ← 可视化路线:Mermaid、phase table、完成度、证据状态
|
|
14
|
+
├── findings.md ← 发现:执行过程中的研究发现和技术决策
|
|
15
|
+
├── progress.md ← 进度:每个阶段的状态更新和验证结果
|
|
16
|
+
├── review.md ← 对抗性审查报告(需要 reviewer / subagent / release review 时必填)
|
|
17
|
+
└── long-running-task-contract.md ← 长程任务合同(仅长程任务需要)
|
|
18
|
+
```
|
|
19
|
+
|
|
20
|
+
复杂任务可以启用 optional structure,但不能默认创建空目录:
|
|
21
|
+
|
|
22
|
+
```
|
|
23
|
+
references/INDEX.md ← 任务本地参考、外部链接、reviewer packet
|
|
24
|
+
artifacts/INDEX.md ← 命令输出、截图、fixture、review transcript
|
|
25
|
+
slices/<slice-id>/brief.md ← 多切片任务的单切片输入和范围
|
|
26
|
+
slices/<slice-id>/evidence.md
|
|
27
|
+
slices/<slice-id>/review.md
|
|
28
|
+
```
|
|
29
|
+
|
|
30
|
+
启用条件:
|
|
31
|
+
|
|
32
|
+
- reviewer/subagent 输入包需要复用:启用 `references/INDEX.md`
|
|
33
|
+
- 命令输出、截图、fixture、review transcript 会污染主文件:启用 `artifacts/INDEX.md`
|
|
34
|
+
- 超过 5 个 slice、多 worker、release gate、L2+ evidence:启用 `slices/`
|
|
35
|
+
|
|
36
|
+
## 执行规则
|
|
37
|
+
|
|
38
|
+
1. **每个阶段前读 task_plan.md** — agent 重新对齐目标
|
|
39
|
+
2. **每个阶段后更新 progress.md** — 记录做了什么、验证了什么
|
|
40
|
+
3. **研究发现写入 findings.md** — 不丢失中间产物
|
|
41
|
+
4. **禁止在项目根目录放过程文件** — task_plan.md、findings.md、progress.md 只能在任务目录内
|
|
42
|
+
5. **对抗性审查必须写 review.md** — 如果任务使用 reviewer / subagent / release review,按 `adversarial-review-standard.md` 写报告
|
|
43
|
+
6. **长程任务必须补合同** — 如果任务需要连续执行、多轮审查或子代理 review,先补 `long-running-task-contract.md`
|
|
44
|
+
7. **任务收口必须回写 Harness Ledger** — 只在任务完成或上下文回写状态变化时记录,不记录每次 `progress.md` 更新
|
|
45
|
+
8. **复杂任务必须记录 `execution_strategy.md`** — 是否使用 subagent、reviewer、worktree、handoff 都写入独立文件。
|
|
46
|
+
9. **非平凡任务必须记录 `visual_roadmap.md`** — HTML dashboard 从独立文件的 Mermaid 和 phase table 计算完成度、阻塞和证据状态。
|
|
47
|
+
10. **路径必须带来源前缀** — 使用 `PUBLIC:`, `PRIVATE:`, `TARGET:`, `EXTERNAL:`, `URL:`,避免脆弱相对路径。
|
|
48
|
+
|
|
49
|
+
## task_plan.md 模板
|
|
50
|
+
|
|
51
|
+
```markdown
|
|
52
|
+
# [任务名称]
|
|
53
|
+
|
|
54
|
+
## 目标
|
|
55
|
+
[一句话说清楚这个任务要达成什么]
|
|
56
|
+
|
|
57
|
+
## 范围
|
|
58
|
+
[做什么、不做什么]
|
|
59
|
+
|
|
60
|
+
## Task IA Budget
|
|
61
|
+
[simple / complex;如果 complex,列出启用哪些 optional structure 和原因]
|
|
62
|
+
|
|
63
|
+
## Context Packet
|
|
64
|
+
| ID | Type | Path | Why It Matters | Used By |
|
|
65
|
+
| --- | --- | --- | --- | --- |
|
|
66
|
+
|
|
67
|
+
## Execution & Visualization Files
|
|
68
|
+
| Contract File | Required | Purpose |
|
|
69
|
+
| --- | --- | --- |
|
|
70
|
+
| `execution_strategy.md` | yes | Execution mode, subagent use, conflict control, evidence depth, handoff rules |
|
|
71
|
+
| `visual_roadmap.md` | yes | Mermaid route, phase table, completion, evidence status, blocking risk |
|
|
72
|
+
|
|
73
|
+
## 步骤
|
|
74
|
+
1. [步骤1]
|
|
75
|
+
2. [步骤2]
|
|
76
|
+
...
|
|
77
|
+
|
|
78
|
+
## 验收标准
|
|
79
|
+
- [ ] [标准1]
|
|
80
|
+
- [ ] [标准2]
|
|
81
|
+
|
|
82
|
+
## Worktree
|
|
83
|
+
- 路径:[worktree 路径]
|
|
84
|
+
- 分支:[分支名]
|
|
85
|
+
- Worker owner:[coordinator / subagent id / 不适用]
|
|
86
|
+
- Worker handoff commit required:[yes / no / 不适用]
|
|
87
|
+
- Coordinator integration branch:[分支名 / 不适用]
|
|
88
|
+
- 若未开 worktree,原因:[说明]
|
|
89
|
+
|
|
90
|
+
## 长程任务判定
|
|
91
|
+
- 是否属于长程任务:[是 / 否]
|
|
92
|
+
- 若是,合同文件:`long-running-task-contract.md`
|
|
93
|
+
- Stop Condition 摘要:[什么时候可以停]
|
|
94
|
+
|
|
95
|
+
## Review 判定
|
|
96
|
+
- 是否需要对抗性 review:[是 / 否]
|
|
97
|
+
- 若是,报告文件:`review.md`
|
|
98
|
+
- Reviewer:[self / subagent / external / human]
|
|
99
|
+
```
|
|
100
|
+
|
|
101
|
+
## 模块并行开发时的任务目录
|
|
102
|
+
|
|
103
|
+
当项目启用模块并行开发(见 `references/module-parallel-standard.md`)时:
|
|
104
|
+
|
|
105
|
+
- 任务目录位于模块内:`docs/09-PLANNING/MODULES/<key>/TASKS/<PREFIX>-NN-<name>/`
|
|
106
|
+
- 跨模块基础设施任务位于:`docs/09-PLANNING/MODULES/_shared/TASKS/YYYY-MM-DD-<name>/`
|
|
107
|
+
- task_plan.md 应填写"模块关联"段(Module、Step、Module Plan link)
|
|
108
|
+
- 会话结束时除了更新 progress.md,还需更新 module_plan.md。
|
|
109
|
+
- 模块 worker 不直接写 Module Registry / Harness Ledger / Closeout SSoT。需要总表同步时,在 task_plan.md 或 progress.md 的 `Coordinator Handoff` 段标记 `pending-coordinator-pass`,由 coordinator 串行同步。
|
|
110
|
+
- coordinator pass 完成后,才更新 Module Registry、Harness Ledger、必要的 Closeout / Regression 表,并把 handoff 标记为 `synced`。
|
|
111
|
+
|
|
112
|
+
## 为什么这套东西有效
|
|
113
|
+
|
|
114
|
+
- agent 的上下文窗口有限,task_plan 是它在长任务中唯一稳定的锚点
|
|
115
|
+
- progress.md 让下一轮 agent(或同一个 agent 的下一个 session)能快速接上
|
|
116
|
+
- findings.md 避免重复研究同一个问题
|
|
117
|
+
- 强制目录结构让所有任务可追溯、可检索
|
|
118
|
+
|
|
119
|
+
## 与 Anthropic Long-running Agents 方案的对照
|
|
120
|
+
|
|
121
|
+
Anthropic 的方案用 Feature List JSON + progress file + git commit 做跨 session 交接。
|
|
122
|
+
task_plan + findings + progress 是同一思路的更细粒度表达;`review.md` 负责保存
|
|
123
|
+
对抗性审查结论,避免 review 只留在对话里。
|
|
124
|
+
|
|
125
|
+
## findings.md 模板
|
|
126
|
+
|
|
127
|
+
```markdown
|
|
128
|
+
# [任务名称] - Findings
|
|
129
|
+
|
|
130
|
+
## 研究发现
|
|
131
|
+
|
|
132
|
+
### [发现主题 1]
|
|
133
|
+
- 背景:[为什么要研究这个]
|
|
134
|
+
- 发现:[具体发现了什么]
|
|
135
|
+
- 影响:[对任务计划有什么影响]
|
|
136
|
+
|
|
137
|
+
### [发现主题 2]
|
|
138
|
+
...
|
|
139
|
+
|
|
140
|
+
## 技术决策
|
|
141
|
+
|
|
142
|
+
| 决策 | 选择 | 原因 | 替代方案 |
|
|
143
|
+
|------|------|------|----------|
|
|
144
|
+
| [决策1] | [选了什么] | [为什么] | [没选什么] |
|
|
145
|
+
```
|
|
146
|
+
|
|
147
|
+
## progress.md 模板
|
|
148
|
+
|
|
149
|
+
```markdown
|
|
150
|
+
# [任务名称] - Progress
|
|
151
|
+
|
|
152
|
+
## 状态:[未开始 / 进行中 / 已完成 / 已阻塞]
|
|
153
|
+
|
|
154
|
+
## 进度记录
|
|
155
|
+
|
|
156
|
+
Evidence values use `type:path:summary`.
|
|
157
|
+
|
|
158
|
+
### [YYYY-MM-DD HH:MM] - [阶段名称]
|
|
159
|
+
- 做了什么:[具体操作]
|
|
160
|
+
- 验证结果:[跑了什么测试,结果如何]
|
|
161
|
+
- 下一步:[接下来做什么]
|
|
162
|
+
- Evidence:[type:path:summary]
|
|
163
|
+
|
|
164
|
+
### [YYYY-MM-DD HH:MM] - [阶段名称]
|
|
165
|
+
...
|
|
166
|
+
|
|
167
|
+
## Residual
|
|
168
|
+
- [遗留问题1]
|
|
169
|
+
- [遗留问题2]
|
|
170
|
+
```
|
|
171
|
+
|
|
172
|
+
## 任务目录命名规范
|
|
173
|
+
|
|
174
|
+
格式:`YYYY-MM-DD-任务名称`
|
|
175
|
+
|
|
176
|
+
示例:
|
|
177
|
+
- `2026-03-15-user-auth-refactor`
|
|
178
|
+
- `2026-03-18-ui-timeline-component`
|
|
179
|
+
- `2026-04-01-regression-gate-webhook-live`
|
|
180
|
+
|
|
181
|
+
## 状态流转
|
|
182
|
+
|
|
183
|
+
```
|
|
184
|
+
未开始 → 进行中 → 已完成
|
|
185
|
+
↓
|
|
186
|
+
已阻塞 → 进行中
|
|
187
|
+
```
|
|
188
|
+
|
|
189
|
+
每次状态变更时,必须同时更新 progress.md 和 Feature SSoT。
|
|
190
|
+
|
|
191
|
+
任务完成时,必须在 `docs/Harness-Ledger.md` 中记录本轮 task plan、SSoT、
|
|
192
|
+
walkthrough、Lessons 检查等上下文回写结果。
|
|
@@ -0,0 +1,167 @@
|
|
|
1
|
+
# 项目诊断(Onboarding Audit)
|
|
2
|
+
|
|
3
|
+
## 目的
|
|
4
|
+
|
|
5
|
+
在搭建 harness 之前,先了解项目现状,确定适合的 harness 规模和结构。
|
|
6
|
+
|
|
7
|
+
## 扫描清单
|
|
8
|
+
|
|
9
|
+
进入项目后,按以下清单逐项检查:
|
|
10
|
+
|
|
11
|
+
### 1. 仓库结构
|
|
12
|
+
- 单仓还是多仓?
|
|
13
|
+
- 是否是 monorepo(多包)?
|
|
14
|
+
- 有哪些子项目 / packages / apps?
|
|
15
|
+
- 有没有前后端分离?
|
|
16
|
+
|
|
17
|
+
### 2. 技术栈
|
|
18
|
+
- 主要语言和框架?
|
|
19
|
+
- 包管理器?(npm / pnpm / yarn / pip / cargo 等)
|
|
20
|
+
- 构建工具?
|
|
21
|
+
- 运行时环境?
|
|
22
|
+
|
|
23
|
+
### 3. 现有文档
|
|
24
|
+
- 有没有 AGENTS.md / CLAUDE.md / COPILOT.md?
|
|
25
|
+
- 有没有 docs/ 目录?结构如何?
|
|
26
|
+
- 有没有 README 以外的开发文档?
|
|
27
|
+
- 有没有架构设计文档?
|
|
28
|
+
|
|
29
|
+
### 4. 现有测试
|
|
30
|
+
- 有没有测试框架?(jest / vitest / pytest / go test 等)
|
|
31
|
+
- 有没有 CI/CD?
|
|
32
|
+
- CI workflow 是否实际存在?路径是什么?
|
|
33
|
+
- PR required checks 是哪些?是否和 workflow job 对齐?
|
|
34
|
+
- 主分支有没有 branch protection?能否用平台 API 验证?
|
|
35
|
+
- 测试覆盖率大概什么水平?
|
|
36
|
+
- 有没有端到端测试或集成测试?
|
|
37
|
+
- 有没有冒烟测试?
|
|
38
|
+
|
|
39
|
+
### 5. 现有任务管理
|
|
40
|
+
- 有没有任务追踪系统?(GitHub Issues / Linear / Jira 等)
|
|
41
|
+
- 有没有排期文档或 SSoT?
|
|
42
|
+
- 有没有 planning 目录或任务模板?
|
|
43
|
+
- 有没有 Harness Ledger 或类似上下文回写总账?
|
|
44
|
+
|
|
45
|
+
### 6. 协作模式
|
|
46
|
+
- 几个人在开发?
|
|
47
|
+
- 是否使用 Coding Agent?用哪些?
|
|
48
|
+
- 是否有多 agent 并行的需求?
|
|
49
|
+
- 是否使用 git worktree?
|
|
50
|
+
- 允许同时存在几个 active worktree?
|
|
51
|
+
- merge 顺序由谁决定?
|
|
52
|
+
|
|
53
|
+
### 6a. Delivery Operating Model
|
|
54
|
+
- 当前是单人主控、多 agent 并行,还是多人团队各自带 agent?
|
|
55
|
+
- 是否有 team lead / tech lead 负责拆 feature block?
|
|
56
|
+
- 是否是前后端分仓、app/service 分仓或 program 多仓?
|
|
57
|
+
- 前端 agent 是否只能看到 API 文档 / mock / schema?后端 agent 是否只能看到消费合同?
|
|
58
|
+
- 当前使用敏捷 sprint、kanban 连续流、瀑布 stage-gate,还是个人连续执行?
|
|
59
|
+
- 是否需要 `docs/09-PLANNING/Delivery-SSoT.md` 记录 feature block owner、依赖、集成顺序和 acceptance gates?
|
|
60
|
+
|
|
61
|
+
### 6b. Repo Governance
|
|
62
|
+
- repo platform 是 GitHub / GitLab / local-only / 其他?
|
|
63
|
+
- 是否有 PR template?
|
|
64
|
+
- 是否有 CODEOWNERS?
|
|
65
|
+
- 是否允许 direct push 到主分支?
|
|
66
|
+
- branch protection 状态是 designed / implemented / verified / blocked-with-owner?
|
|
67
|
+
- agent 是否有权限读取或设置 repo protection?
|
|
68
|
+
|
|
69
|
+
### 7. 关键 Surface
|
|
70
|
+
- 项目有哪些用户入口?(Web UI、API、CLI、Bot、插件等)
|
|
71
|
+
- 有哪些外部集成点?(第三方 API、数据库、消息队列等)
|
|
72
|
+
- 哪些 surface 最容易被改动破坏?
|
|
73
|
+
|
|
74
|
+
## 诊断报告模板
|
|
75
|
+
|
|
76
|
+
扫描完成后,输出以下格式的诊断报告:
|
|
77
|
+
|
|
78
|
+
```markdown
|
|
79
|
+
# Harness Onboarding Audit
|
|
80
|
+
|
|
81
|
+
## 项目概况
|
|
82
|
+
- 项目名:[名称]
|
|
83
|
+
- 仓库类型:[单仓 / monorepo / 多仓]
|
|
84
|
+
- 技术栈:[语言 / 框架 / 运行时]
|
|
85
|
+
- 团队规模:[人数 + agent 数]
|
|
86
|
+
|
|
87
|
+
## 现状评估
|
|
88
|
+
|
|
89
|
+
| 维度 | 现状 | 评级 |
|
|
90
|
+
|------|------|------|
|
|
91
|
+
| AGENTS.md | [有/无/需改造] | 🟢/🟡/🔴 |
|
|
92
|
+
| docs/ 目录 | [有/无/需改造] | 🟢/🟡/🔴 |
|
|
93
|
+
| Reference 标准 | [有/无/需改造] | 🟢/🟡/🔴 |
|
|
94
|
+
| Planning Loop | [有/无/需改造] | 🟢/🟡/🔴 |
|
|
95
|
+
| Delivery Operating Model | [solo/team/split-repo/program/waterfall/kanban/需确认] | 🟢/🟡/🔴 |
|
|
96
|
+
| Delivery SSoT | [有/无/不需要/需改造] | 🟢/🟡/🔴 |
|
|
97
|
+
| Feature SSoT | [有/无/需改造] | 🟢/🟡/🔴 |
|
|
98
|
+
| Regression 体系 | [有/无/需改造] | 🟢/🟡/🔴 |
|
|
99
|
+
| CI/CD | [有/无/需改造] | 🟢/🟡/🔴 |
|
|
100
|
+
| Repo Governance | [有/无/需改造] | 🟢/🟡/🔴 |
|
|
101
|
+
| Branch Protection | [designed/implemented/verified/blocked] | 🟢/🟡/🔴 |
|
|
102
|
+
| Required Checks | [有/无/需改造] | 🟢/🟡/🔴 |
|
|
103
|
+
| Harness Ledger | [有/无/需改造] | 🟢/🟡/🔴 |
|
|
104
|
+
| Walkthrough 流程 | [有/无/需改造] | 🟢/🟡/🔴 |
|
|
105
|
+
| Worktree 规范 | [有/无/需改造] | 🟢/🟡/🔴 |
|
|
106
|
+
|
|
107
|
+
## 关键 Surface 清单
|
|
108
|
+
1. [Surface 1]:[描述]
|
|
109
|
+
2. [Surface 2]:[描述]
|
|
110
|
+
...
|
|
111
|
+
|
|
112
|
+
## 推荐 Harness 规模
|
|
113
|
+
[Lite / Standard / Full](见下方项目类型分支)
|
|
114
|
+
|
|
115
|
+
## 落地方案
|
|
116
|
+
[具体说明需要创建哪些文件、改造哪些现有文件]
|
|
117
|
+
|
|
118
|
+
## 风险点
|
|
119
|
+
- [风险1]
|
|
120
|
+
- [风险2]
|
|
121
|
+
```
|
|
122
|
+
|
|
123
|
+
## 项目类型分支
|
|
124
|
+
|
|
125
|
+
根据项目规模和复杂度,harness 分三个规模:
|
|
126
|
+
|
|
127
|
+
### Lite(小型项目)
|
|
128
|
+
适用于:单仓、单人开发、代码量 < 1 万行、surface 少于 3 个
|
|
129
|
+
|
|
130
|
+
最小配置:
|
|
131
|
+
- AGENTS.md
|
|
132
|
+
- docs/11-REFERENCE/ 下 2-3 个标准文件
|
|
133
|
+
- Delivery Operating Model 标准,明确是否为 `solo-orchestrator`
|
|
134
|
+
- Planning task plan / findings / progress / review 模板
|
|
135
|
+
- repo governance / CI-CD 标准和 residual
|
|
136
|
+
- 简化版 Regression SSoT(可以只有 tests + local_smoke 两层)
|
|
137
|
+
- Harness Ledger(只记录 task closeout 行)
|
|
138
|
+
- Walkthrough 模板
|
|
139
|
+
|
|
140
|
+
可省略:
|
|
141
|
+
- Cadence Ledger(手动触发即可)
|
|
142
|
+
- Feature SSoT(用 AGENTS.md 里的简单列表替代)
|
|
143
|
+
- Worktree 规范(单人不需要并行)
|
|
144
|
+
|
|
145
|
+
### Standard(中型项目)
|
|
146
|
+
适用于:单仓或 monorepo、1-3 人 + agent、代码量 1-10 万行、surface 3-10 个
|
|
147
|
+
|
|
148
|
+
完整配置:
|
|
149
|
+
- 全部 Phase 1-12
|
|
150
|
+
- Delivery Operating Model;若多人协作则创建 Delivery SSoT
|
|
151
|
+
- Evidence Depth 至少覆盖到 L3(live 环境验证)
|
|
152
|
+
- Cadence Ledger
|
|
153
|
+
- Harness Ledger
|
|
154
|
+
- Worktree 规范
|
|
155
|
+
- repo governance / CI-CD workflow 或 blocked-with-owner residual
|
|
156
|
+
|
|
157
|
+
### Full(大型项目)
|
|
158
|
+
适用于:多仓或大型 monorepo、多人 + 多 agent 并行、代码量 > 10 万行、surface > 10 个
|
|
159
|
+
|
|
160
|
+
完整配置 + 额外要求:
|
|
161
|
+
- Program / split-repo operating model 和 Delivery SSoT
|
|
162
|
+
- 每个子仓库或重要子包有自己的 reference 文件
|
|
163
|
+
- Evidence Depth 要求覆盖到 L4 或 L5
|
|
164
|
+
- Shared Regression Batch 定期执行
|
|
165
|
+
- Harness Ledger 季度归档
|
|
166
|
+
- 多 agent 并行分工协议
|
|
167
|
+
- 跨仓库 surface 映射
|
|
@@ -0,0 +1,89 @@
|
|
|
1
|
+
# Regression 体系
|
|
2
|
+
|
|
3
|
+
## 核心思路
|
|
4
|
+
|
|
5
|
+
单元测试只是底线。长程项目需要多层证据来保证正确性。
|
|
6
|
+
|
|
7
|
+
## Evidence Depth 五级制
|
|
8
|
+
|
|
9
|
+
每条回归面都标注它的证据到了哪一层:
|
|
10
|
+
|
|
11
|
+
| 层级 | 名称 | 含义 | 可信度 |
|
|
12
|
+
|------|------|------|--------|
|
|
13
|
+
| L1 | tests | 只有单元测试 | 最低 |
|
|
14
|
+
| L2 | local_smoke | 本地冒烟测试通过 | 低 |
|
|
15
|
+
| L3 | live_e2e | 真实环境端到端验证 | 中 |
|
|
16
|
+
| L4 | browser_human_proxy | 浏览器模拟真人操作 | 高 |
|
|
17
|
+
| L5 | hard_gate | 结构化判定 + 非零退出 | 最高 |
|
|
18
|
+
|
|
19
|
+
越高层的证据越可信。L1 只能说明代码能编译和通过基本逻辑检查,L5 能自动告诉你"过了"还是"没过"。
|
|
20
|
+
|
|
21
|
+
## Regression SSoT 结构
|
|
22
|
+
|
|
23
|
+
```markdown
|
|
24
|
+
# Regression SSoT
|
|
25
|
+
|
|
26
|
+
## Active Fixed Gates
|
|
27
|
+
|
|
28
|
+
| ID | Status | Surface | Primary Entrypoint | Evidence Depth |
|
|
29
|
+
|----|--------|---------|-------------------|----------------|
|
|
30
|
+
| RG-001 | 🟢 | API Contract Smoke | `npm run smoke:api` | hard_gate |
|
|
31
|
+
| RG-002 | 🟢 | External Integration E2E | `npm run smoke:integration:live` | live_e2e |
|
|
32
|
+
| ...
|
|
33
|
+
|
|
34
|
+
## Residual Items
|
|
35
|
+
|
|
36
|
+
| ID | Surface | Issue | Priority |
|
|
37
|
+
|----|---------|-------|----------|
|
|
38
|
+
| R-001 | Frontend | Timeline 组件偶发渲染延迟 | P2 |
|
|
39
|
+
| ...
|
|
40
|
+
|
|
41
|
+
## Shared Regression Ledger
|
|
42
|
+
|
|
43
|
+
| Batch | Date | Scope | Result | Next Checkpoint |
|
|
44
|
+
|-------|------|-------|--------|-----------------|
|
|
45
|
+
| SRB-010 | 2026-04-12 | Full | 9/9 🟢 | SRB-011 after wave-14 |
|
|
46
|
+
| ...
|
|
47
|
+
|
|
48
|
+
## Archive Index
|
|
49
|
+
|
|
50
|
+
| Archive | Range | Notes |
|
|
51
|
+
|---------|-------|-------|
|
|
52
|
+
| `docs/05-TEST-QA/_archive/Regression-SSoT-archive-YYYY-QN.md` | RG-XXX-RG-YYY | retired gates |
|
|
53
|
+
```
|
|
54
|
+
|
|
55
|
+
## Cadence Ledger
|
|
56
|
+
|
|
57
|
+
定义什么情况下自动触发哪些回归面:
|
|
58
|
+
|
|
59
|
+
```markdown
|
|
60
|
+
## Cadence Rules
|
|
61
|
+
|
|
62
|
+
- 改了 API contract → 跑 RG-001 + RG-004
|
|
63
|
+
- 改了 external integration adapter → 跑 RG-002 + RG-003
|
|
64
|
+
- 改了 core domain logic → 跑 RG-008 + RG-009 + RG-010
|
|
65
|
+
- 改了 frontend user flow → 跑 RG-005(如有)
|
|
66
|
+
- 任何 merge 到 master → 跑 Full Shared Batch
|
|
67
|
+
```
|
|
68
|
+
|
|
69
|
+
不用人记住该跑什么,系统自己知道。
|
|
70
|
+
|
|
71
|
+
## 与 Harness Ledger 的关系
|
|
72
|
+
|
|
73
|
+
Regression SSoT 和 Cadence Ledger 管“回归事实”:有哪些 gate、证据深度是多少、
|
|
74
|
+
最近 batch 结果如何。
|
|
75
|
+
|
|
76
|
+
Harness Ledger 只在当前任务收口时记录 `Regression=updated`、`Regression=n/a` 等
|
|
77
|
+
任务级回写状态。不要把 batch 明细复制进 Harness Ledger。
|
|
78
|
+
|
|
79
|
+
## 建立回归体系的步骤
|
|
80
|
+
|
|
81
|
+
1. 列出项目的所有关键 surface(用户入口、API 端点、集成点)
|
|
82
|
+
2. 为每个 surface 建一条 regression gate,写好命令行入口
|
|
83
|
+
3. 标注每条 gate 的当前 Evidence Depth
|
|
84
|
+
4. 定义 Cadence Rules
|
|
85
|
+
5. 跑第一轮 Shared Batch,记录结果
|
|
86
|
+
6. 持续迭代:每次新增 surface 或 evidence depth 提升时更新 SSoT
|
|
87
|
+
|
|
88
|
+
废弃 gate 或过期 shared batch 明细不要长期堆在 Active 表底部;移动到
|
|
89
|
+
`docs/05-TEST-QA/_archive/`,并在 Regression SSoT 中保留 Archive Index。
|
|
@@ -0,0 +1,131 @@
|
|
|
1
|
+
# Repository Governance Standard
|
|
2
|
+
|
|
3
|
+
## 核心思路
|
|
4
|
+
|
|
5
|
+
Harness 不是只生成文档骨架。每个项目必须有项目级 repository governance contract,
|
|
6
|
+
明确分支、PR、merge、review、required checks、worktree 并发和权限残项。
|
|
7
|
+
|
|
8
|
+
如果某些设置因为权限不足无法自动配置,agent 不能把它们当作完成;必须记录为
|
|
9
|
+
`blocked-with-owner` 或 `manual-setup-residual`。
|
|
10
|
+
|
|
11
|
+
## 存放位置
|
|
12
|
+
|
|
13
|
+
标准文件:
|
|
14
|
+
|
|
15
|
+
```text
|
|
16
|
+
docs/11-REFERENCE/repo-governance-standard.md
|
|
17
|
+
```
|
|
18
|
+
|
|
19
|
+
如果项目需要更详细的 CI/CD 说明,另见:
|
|
20
|
+
|
|
21
|
+
```text
|
|
22
|
+
docs/11-REFERENCE/ci-cd-standard.md
|
|
23
|
+
```
|
|
24
|
+
|
|
25
|
+
## 必填字段
|
|
26
|
+
|
|
27
|
+
每个项目必须定制以下内容:
|
|
28
|
+
|
|
29
|
+
### Repo Platform Profile
|
|
30
|
+
|
|
31
|
+
- Platform: [GitHub / GitLab / local-only / other]
|
|
32
|
+
- Remote: [owner/repo or URL]
|
|
33
|
+
- Default branch: [main / master / other]
|
|
34
|
+
- Repo type: [single app / monorepo / multi-repo / library / service]
|
|
35
|
+
- Admin access available to agent: [yes / no / unknown]
|
|
36
|
+
|
|
37
|
+
### Branch Model
|
|
38
|
+
|
|
39
|
+
- Protected branch(es)
|
|
40
|
+
- Feature branch naming
|
|
41
|
+
- Release branch naming, if any
|
|
42
|
+
- Hotfix branch naming, if any
|
|
43
|
+
- Direct push policy
|
|
44
|
+
|
|
45
|
+
### PR Policy
|
|
46
|
+
|
|
47
|
+
- PR required before merge: [yes / no]
|
|
48
|
+
- PR title format
|
|
49
|
+
- PR body requirements
|
|
50
|
+
- Required reviewers
|
|
51
|
+
- Required review type: [self / subagent / external / human]
|
|
52
|
+
- Merge method: [squash / merge commit / rebase]
|
|
53
|
+
- Who decides merge order
|
|
54
|
+
|
|
55
|
+
### Required Checks
|
|
56
|
+
|
|
57
|
+
List every check required before merge:
|
|
58
|
+
|
|
59
|
+
| Check | Command / Workflow | Required? | Evidence |
|
|
60
|
+
|-------|--------------------|-----------|----------|
|
|
61
|
+
| lint | [command] | yes/no | [where result is recorded] |
|
|
62
|
+
| typecheck | [command] | yes/no | [where result is recorded] |
|
|
63
|
+
| build | [command] | yes/no | [where result is recorded] |
|
|
64
|
+
| test | [command] | yes/no | [where result is recorded] |
|
|
65
|
+
| smoke | [command] | yes/no | [where result is recorded] |
|
|
66
|
+
|
|
67
|
+
### Branch Protection Plan
|
|
68
|
+
|
|
69
|
+
Branch protection status must use one of:
|
|
70
|
+
|
|
71
|
+
- `designed`
|
|
72
|
+
- `implemented`
|
|
73
|
+
- `verified`
|
|
74
|
+
- `blocked-with-owner`
|
|
75
|
+
|
|
76
|
+
Required fields:
|
|
77
|
+
|
|
78
|
+
- Required status checks
|
|
79
|
+
- Required PR review count
|
|
80
|
+
- Dismiss stale reviews: [yes / no]
|
|
81
|
+
- Require branches up to date: [yes / no]
|
|
82
|
+
- Block force push: [yes / no]
|
|
83
|
+
- Block deletion: [yes / no]
|
|
84
|
+
- Bypass actors, if any
|
|
85
|
+
- Verification command or manual setup residual
|
|
86
|
+
|
|
87
|
+
### Worktree Concurrency
|
|
88
|
+
|
|
89
|
+
- Max active worktrees
|
|
90
|
+
- Naming pattern
|
|
91
|
+
- Branch pattern
|
|
92
|
+
- Ownership rule
|
|
93
|
+
- Subagent worker rule: each code-changing worker uses its own worktree / branch and hands off a commit SHA
|
|
94
|
+
- Merge ordering rule
|
|
95
|
+
- Cleanup rule
|
|
96
|
+
|
|
97
|
+
## Evidence Status Model
|
|
98
|
+
|
|
99
|
+
Every governance item must be marked as:
|
|
100
|
+
|
|
101
|
+
| Status | Meaning |
|
|
102
|
+
|--------|---------|
|
|
103
|
+
| `designed` | Plan exists, not implemented |
|
|
104
|
+
| `implemented` | File / workflow / config exists |
|
|
105
|
+
| `verified` | Live or local verification passed |
|
|
106
|
+
| `blocked-with-owner` | Cannot finish without named owner/action |
|
|
107
|
+
|
|
108
|
+
Agent must not describe `designed` as complete.
|
|
109
|
+
|
|
110
|
+
## GitHub Default Adapter
|
|
111
|
+
|
|
112
|
+
For GitHub repositories, the default implementation should include:
|
|
113
|
+
|
|
114
|
+
- `.github/pull_request_template.md`
|
|
115
|
+
- `.github/workflows/ci.yml` or an explicit reason why CI is impossible
|
|
116
|
+
- branch protection plan for `main`
|
|
117
|
+
- required checks matching actual workflow job names
|
|
118
|
+
- review routing rule aligned with `review-routing-standard.md`
|
|
119
|
+
|
|
120
|
+
If the agent has GitHub admin permissions, it should verify branch protection with `gh api`.
|
|
121
|
+
If not, it must write manual setup residual with owner and exact settings.
|
|
122
|
+
|
|
123
|
+
## Completion Rule
|
|
124
|
+
|
|
125
|
+
Bootstrap is not complete unless repository governance is at least:
|
|
126
|
+
|
|
127
|
+
- PR policy: `implemented`
|
|
128
|
+
- Required checks: `implemented`
|
|
129
|
+
- Branch protection: `designed` with residual, or `verified`
|
|
130
|
+
- Worktree concurrency: `implemented`
|
|
131
|
+
- Harness checker: passing or blocked-with-owner with explicit residual
|
|
@@ -0,0 +1,103 @@
|
|
|
1
|
+
# Review Routing Standard
|
|
2
|
+
|
|
3
|
+
## 核心思路
|
|
4
|
+
|
|
5
|
+
`adversarial-review-standard.md` 定义 review report 怎么写;本标准定义 review 怎么触发、谁来审、外部 reviewer 如何纳入项目规则。
|
|
6
|
+
|
|
7
|
+
每个 planned task / wave / feature 结束前,必须自动进入 closeout review。默认最低要求是:
|
|
8
|
+
|
|
9
|
+
1. 主 agent 完成 self-review。
|
|
10
|
+
2. 主 agent 调用 reviewer / subagent 做外部视角审查。
|
|
11
|
+
3. reviewer 按 `adversarial-review-standard.md` 写入或补全 `review.md`。
|
|
12
|
+
4. 主 agent 修复或路由 material findings。
|
|
13
|
+
5. 再次运行 Confidence Challenge,直到没有 open material finding。
|
|
14
|
+
|
|
15
|
+
## 触发规则
|
|
16
|
+
|
|
17
|
+
必须触发 closeout review:
|
|
18
|
+
|
|
19
|
+
- planned task / wave / feature 收口
|
|
20
|
+
- 长程任务每轮进入 stop condition 判断前
|
|
21
|
+
- 涉及架构、数据、安全、权限、部署、迁移、跨模块契约
|
|
22
|
+
- release / PR / merge 前
|
|
23
|
+
- 用户明确要求 review / 外部审查 / 人工审查
|
|
24
|
+
|
|
25
|
+
轻量单文件修复可以只做 self-review,但必须在 `progress.md` 写明跳过 subagent / external reviewer 的理由。
|
|
26
|
+
|
|
27
|
+
## Reviewer 层级
|
|
28
|
+
|
|
29
|
+
| 层级 | Reviewer | 适用场景 | 要求 |
|
|
30
|
+
|------|----------|----------|------|
|
|
31
|
+
| L0 | Self-review | 微小变更 | 写入 `progress.md` 或 `review.md` |
|
|
32
|
+
| L1 | Subagent reviewer | 默认非平凡任务 | 必须写 `review.md` |
|
|
33
|
+
| L2 | External agent reviewer | 用户或项目要求,如 Claude Code / Gemini / Codex 另一实例 | 必须写 `review.md`,并记录 reviewer identity |
|
|
34
|
+
| L3 | Human reviewer | 高风险产品、架构、安全、数据、发布判断 | Agent 必须在内部审查结束后明确询问是否需要人工审查 |
|
|
35
|
+
|
|
36
|
+
默认策略:planned task 至少 L1。若当前环境无法调用 subagent,必须记录
|
|
37
|
+
`skipped-with-reason`,并升级 self-review 的 Confidence Challenge 严格度。
|
|
38
|
+
|
|
39
|
+
## Subagent Worker Routing
|
|
40
|
+
|
|
41
|
+
本标准默认把 subagent 当 reviewer:只读审查、写 `review.md`、报告 material findings。
|
|
42
|
+
|
|
43
|
+
如果 subagent 被要求直接改代码、测试、产品文档或 harness 文档,它就不是 reviewer,
|
|
44
|
+
而是 worker。Worker 必须按 `worktree-parallel.md` / 项目级 `worktree-standard.md`
|
|
45
|
+
执行:
|
|
46
|
+
|
|
47
|
+
- coordinator 先分配独立 worktree / branch、任务目录和 write scope
|
|
48
|
+
- worker 只在自己的 worktree 内实现、验证并提交
|
|
49
|
+
- handoff 必须包含 worktree path、branch、commit SHA、checks、residual risks
|
|
50
|
+
- coordinator 负责 merge / conflict resolution / final gates
|
|
51
|
+
|
|
52
|
+
禁止把多个 worker 的未提交改动混在 coordinator 当前 checkout,再由 coordinator 一次性提交。
|
|
53
|
+
|
|
54
|
+
## 外部审查人工触发
|
|
55
|
+
|
|
56
|
+
如果用户要求外部审查或人工审查,Agent 不应把这当成一次聊天请求。它应转化为项目规则:
|
|
57
|
+
|
|
58
|
+
1. 在当前任务的 `review.md` 记录触发来源和审查范围。
|
|
59
|
+
2. 如果是一次性审查,按本轮任务执行。
|
|
60
|
+
3. 如果用户希望长期遵循,创建或更新项目级 `docs/11-REFERENCE/review-routing-standard.md`。
|
|
61
|
+
4. 在 `AGENTS.md` 的 Task-Type Reading Matrix 中加入 reviewer routing 入口。
|
|
62
|
+
5. 在 Harness Ledger 记录该规则变更。
|
|
63
|
+
|
|
64
|
+
## 项目级 reviewer policy
|
|
65
|
+
|
|
66
|
+
项目可以指定默认外部 reviewer,例如:
|
|
67
|
+
|
|
68
|
+
- Claude Code 负责架构和实现策略外部审查
|
|
69
|
+
- Gemini 负责 web / dependency / ecosystem research 交叉检查
|
|
70
|
+
- Codex subagent 负责代码 diff、测试缺口和回归风险
|
|
71
|
+
- Human reviewer 负责产品方向、安全和发布 gate
|
|
72
|
+
|
|
73
|
+
项目级 policy 必须写清:
|
|
74
|
+
|
|
75
|
+
- 默认 reviewer
|
|
76
|
+
- 触发条件
|
|
77
|
+
- reviewer 是否只读
|
|
78
|
+
- reviewer 输出写入位置
|
|
79
|
+
- 主 agent 如何处理分歧
|
|
80
|
+
- 哪些 finding 必须暂停并询问用户
|
|
81
|
+
|
|
82
|
+
## 两方校准
|
|
83
|
+
|
|
84
|
+
当引入新的外部 reviewer 时,应先执行一次 calibration:
|
|
85
|
+
|
|
86
|
+
1. 主 agent 写出任务目标、scope、证据和当前策略。
|
|
87
|
+
2. 外部 reviewer 用 Confidence Challenge 挑战策略。
|
|
88
|
+
3. 双方对 finding severity、accepted residual、stop condition 口径达成一致。
|
|
89
|
+
4. 将共识写入项目级 `review-routing-standard.md`。
|
|
90
|
+
|
|
91
|
+
校准不是长期讨论。它的目标是形成后续可执行规则。
|
|
92
|
+
|
|
93
|
+
## Closeout Checklist
|
|
94
|
+
|
|
95
|
+
任务收口前必须确认:
|
|
96
|
+
|
|
97
|
+
- [ ] `review.md` 存在,或有明确 `skipped-with-reason`
|
|
98
|
+
- [ ] L1 subagent review 已执行,或环境限制已记录
|
|
99
|
+
- [ ] Confidence Challenge 已回答并记录 final confidence basis
|
|
100
|
+
- [ ] open P0/P1 findings 为 0
|
|
101
|
+
- [ ] material P2 已修复或 accepted residual 并路由
|
|
102
|
+
- [ ] 如使用 worker subagent,已记录 worker branch、commit SHA、checks 和 integration evidence
|
|
103
|
+
- [ ] walkthrough / Harness Ledger 引用了 review report 或 skip reason
|