project-audit
Skill0 starsChanged 4 months agoMITView on GitHub
Use when auditing a project for completeness, consistency, and readiness — "audit project", "项目审计", "check project", "verify project", "检查项目", "review project". This skill checks directory structure, file completeness, navigation consistency, cross-file coherence, traceability, standards compliance, and phase compliance. Different from stage-review (single stage) — this is a whole-project audit. Can be invoked with depth: quick/standard/deep.
15 other projects use a version of this.
Adopted by
Most recently taken up by
- blamechris/chroxytheir own version, first seen Nov 8, 2026
- AhmedAbdelfattah0/AI-Skillstheir own version, first seen Nov 4, 2026
- grooownow/qa-pilottheir own version, first seen Nov 1, 2026
- hmj1026/dhpktheir own version, first seen Nov 1, 2026
- Kin9Zeus/senior-engineer-skillstheir own version, first seen Nov 1, 2026
- Khaymah/sabry-claude-kittheir own version, first seen Nov 1, 2026
Used with
Projects with it also use merge-prep, repo-intake and 2 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
sd0xdevSkillmerge-prep
Pre-merge analysis and preparation. Analyzes source branch vs target branch: commit stats, conflict detection, file impact. Analysis-only v1 — outputs report + suggested commands, does not auto-merge. Use when: user says 'merge prep', 'pre-merge', 'merge analysis', or /merge-prep
Also in 3 projects that use this
AuroraBackcountrySkillrepo-intake
Get a picked-up repo ready to work in. Unfamiliar repo: ask intent, map it by reading, write the canonical AGENTS.md, wire the vault, propose a first task. Already-documented repo with no vault space: wire-only — connect it and leave the docs alone. Use when picking up a codebase: "onboard me", "get me set up on this repo", "I cloned/forked this", "map this repo", "wire this into the vault" — or proactively when a repo has no AGENTS.md/CLAUDE.md and real work starts. Check ownership first: on a repo the user doesn't own, intake is vault-first and writes nothing into tracked files. NOT memory refresh (that's memory-checkpoint) or greenfield scaffolding (/scaffold); documented AND wired = nothing to do.
Also in 3 projects that use this
sd0xdevSkillissue-analyze
GitHub Issue and PR review thread deep analysis with Codex blind verdict. Use when: analyzing issue root cause, classifying problems, investigation planning, triaging PR review comments for actionability. Not for: fixing bugs (use bug-fix), code exploration (use code-explore). Output: classified analysis + verdict assessment + investigation strategy.
Also in 3 projects that use this
apacheSkillprecommit
Validate and fix Pinot pre-commit checks on affected modules before a commit or PR.
Also in 3 projects that use this
Other versions15 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 15
- sd0xdev/sd0x-dev-flow★ 192same name
- sd0xdev/sd0x-harness★ 192same name
- danmaps/gis-agent-skills★ 30same name
- NeuralTrust/TrustGate★ 17same name
- blamechris/chroxy★ 8same name
- grooownow/qa-pilot★ 6same name
- AhmedAbdelfattah0/AI-Skills★ 4same name
- Kin9Zeus/senior-engineer-skills★ 3same name
- hmj1026/dhpk★ 2same name
- DentalEdge-Solutions/claude_code★ 1same name
5 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository34 other files it gives its agents.
- AGENTS.mdAGENTS.md
- acceptance-designSkill
- adversarial-scenariosSkill
- architecture-decisionsSkill
- assumptions-trackingSkill
- competitive-analysisSkill
- consistency-reviewSkill
- copilot-session-retroSkill
- data-api-designSkill
- decision-logSkill
- delivery-planningSkill
- execution-logSkill
- express-incubateSkill
- history-syncSkill
- home-navigationSkill
- implementation-kitSkill
- incubation-discoverySkill
- opencode-session-retroSkill
- operations-planningSkill
- overview-framingSkill
- performance-designSkill
- plain-language-handoffSkill
- project-bootstrapSkill
- prompt-authoringSkill
- quantified-acceptanceSkill
- requirements-specSkill
- research-spikesSkill
- retrospectiveSkill
- risk-registerSkill
- security-designSkill
- skill-creatorSkill
- stage-reviewSkill
- test-strategySkill
- ui-designSkill
The file
Files
Project Audit
全项目级别的结构和内容审计。不同于 stage-review(单阶段文档质量),project-audit 检查跨文件、跨目录的全量一致性。
定位
| 维度 | 说明 |
| 目标 | 发现项目中结构不全、内容矛盾、链接断裂、规范违规、重复条目、阶段标记过期 |
| 输入 | 项目根目录 |
| 输出 | 结构化审计报告(JSON + Markdown) |
| 与 stage-review 区别 | stage-review 验单文档质量;project-audit 验整体一致性 |
When to Use
- 快速孵化完成后(
express-incubate末尾) - 阶段切换前(需求→架构、架构→交付)
- 生成
implementation-kit之前 - 用户说"检查项目"、"审计"、"verify"、"review project"
- 用户说"分析孵化情况"、"看看还有什么缺的"
Input
| 来源 | 内容 | 必读 |
| 项目目录 | <项目根目录>/ 完整递归文件列表 | ✅ |
| 模板参考 | 仓库根目录 ../../../PROJECT_WIKI_TEMPLATE.md(唯一标准) | ✅ |
| HOME.md | 首页导航、阶段标记、决策摘要、风险摘要 | ✅ |
| 所有 wiki 文件 | 逐文件读取内容 | deep 模式 ✅ |
Output
Output Schema
{
"project": "<项目名>",
"timestamp": "ISO 8601",
"depth": "quick|standard|deep",
"score": 75,
"verdict": "PASS|WARNING|NEEDS_FIX|BLOCKED",
"categories": {
"structure": {"status": "PASS|WARNING|BLOCKER", "score": 0-100, "issues": [...]},
"files": {"status": "...", "score": 0-100, "issues": [...]},
"navigation": {"status": "...", "score": 0-100, "issues": [...]},
"coherence": {"status": "...", "score": 0-100, "issues": [...]},
"traceability": {"status": "...", "score": 0-100, "issues": [...]},
"compliance": {"status": "...", "score": 0-100, "issues": [...]},
"phase": {"status": "...", "score": 0-100, "issues": [...]}
},
"findings": {
"blockers": [{"id": "F1", "desc": "...", "fix": "..."}],
"warnings": [{"id": "C5", "desc": "...", "fix": "..."}],
"infos": [{"id": "N2", "desc": "...", "suggestion": "..."}]
}
}Minimum Viable Output
最少包含:
- 7 个维度的 pass/warning/blocker 判定
- 每个异常的检查项 ID、描述、修复建议
- 总分和 verdict
Complete Output
完整输出包含:
- Minimum Viable Output 全部内容
- 每个异常的具体文件路径和行号
- 深度模式下逐文件内容质量评定
- 修复优先级排序
reports/audit/<timestamp>.json+reports/audit/<timestamp>.md
文件产出
| 文件 | 说明 |
reports/audit/<timestamp>.json | 结构化审计数据 |
reports/audit/<timestamp>.md | 人类可读审计报告 |
Dependencies
| 类型 | 说明 |
| 前置 | 项目至少完成快速孵化(有 HOME.md + 最小文件集) |
| 后置 | 修复审计问题后可进入下一阶段 |
| 依赖读取 | ../../../PROJECT_WIKI_TEMPLATE.md(仓库唯一模板标准) |
Procedure
Phase 1: 确定审计深度
| 深度 | 检查范围 | 适用场景 |
quick | ①结构(S1-S4) + ②文件(F1-F5) + ③导航(N1) + ⑦阶段(P1-P2) | express-incubate 后快速扫一眼 |
standard | 全部 7 维度所有检查项 | 阶段切换、kit 生成前 |
deep | standard + 逐文件内容质量分析 + 统一页面骨架合规 | 验收/交接前 |
Phase 2: 7 维度逐项检查
① 结构完整性(Structure)
检查实际目录结构 vs 模板结构。
| ID | 检查项 | 判定 |
| S1 | 所有模板目录存在 | 缺失目录 = 🟡 WARNING |
| S2 | 目录中至少有一个文件 | 完全空目录 = 🟡 WARNING(除非有 README 声明跳过) |
| S3 | 无无关目录/文件 | 非模板目录 = 🟢 INFO |
| S4 | 任务 ID 无重复(task-breakdown.md 和 backlog.md) | 重复 ID = 🔴 BLOCKER |
express-incubate 特例:以下目录在快速模式下允许为空的(但应有 README 说明):
wiki/02-research/、wiki/04-data-and-api/、wiki/06-quality/、wiki/07-operations/reports/子目录
② 文件完整性(Files)
检查各目录下的文件是否齐全。
| ID | 检查项 | 判定 |
| F1 | 最小文件集存在 | 缺失最小文件 = 🔴 BLOCKER |
| F2 | task-breakdown.md 与 backlog.md 任务 ID 一致 | 不一致 = 🔴 BLOCKER |
| F3 | implementation-kit 的 prompt 数与 task 数一致 | 不一致 = 🟡 WARNING |
| F4 | implementation-kit 的 acceptance 数与 task 数一致 | 不一致 = 🟡 WARNING |
| F5 | tasks/ 目录完整:backlog.md + in-progress.md + done.md | 缺 in-progress 或 done = 🟡 WARNING |
| F6 | 阶段相关关键交付物存在(见下表) | 缺失 = 🟡 WARNING |
| F7 | reports/ 子目录有实际内容(非空目录) | 空目录 = 🟢 INFO |
最小文件集(即使快速模式也应存在):
HOME.md,wiki/00-overview/vision.md,wiki/00-overview/scope.mdwiki/01-requirements/user-stories.md,wiki/01-requirements/constraints.mdwiki/01-requirements/assumptions.md,wiki/01-requirements/open-questions.mdwiki/03-architecture/system-design.mdwiki/05-delivery/task-breakdown.md,tasks/backlog.md
F6 阶段-交付物映射:
| 项目阶段 | 期望存在的关键交付物 | 缺失判定 |
| 需求澄清后 | user-stories.md、constraints.md、assumptions.md | 🟡 WARNING |
| 调研验证后 | wiki/02-research/technical-spikes.md、wiki/02-research/risk-register.md | 🟡 WARNING |
| 架构设计后 | system-design.md(≥3 模块)、tech-stack.md | 🟡 WARNING |
| 一致性审查后 | wiki/06-quality/consistency-report.md(高严重矛盾已处理) | 🟡 WARNING |
| 设计并行批次后 | ui-design.md、security.md、performance.md、test-strategy.md、threat-scenarios.md | 🟡 WARNING |
| 交付规划后 | task-breakdown.md(≥5 任务)、tasks/in-progress.md、tasks/done.md | 🟡 WARNING |
| 设计验收后 | wiki/06-quality/acceptance-plan.md、wiki/06-quality/quantified-checklist.md | 🟡 WARNING |
| 运营规划后 | wiki/07-operations/deployment.md | 🟡 WARNING |
| implementation-kit 存在 | 说明至少完成架构设计,应检查上述所有阶段交付物 | 加权判定 |
③ 导航一致性(Navigation)
| ID | 检查项 | 判定 |
| N1 | HOME.md 所有链接目标文件存在 | 断裂链接 = 🔴 BLOCKER |
| N2 | HOME.md 链接覆盖所有非空 wiki 目录 | 遗漏 = 🟡 WARNING |
| N3 | wiki 间交叉引用有效 | 断裂 = 🟡 WARNING |
N2 增强:如果项目已生成 implementation-kit,HOME.md 导航至少应覆盖:
wiki/02-research/risk-register.md(风险登记)wiki/06-quality/acceptance-plan.md(验收计划)wiki/07-operations/deployment.md(部署说明)
④ 内容一致性(Coherence)
| ID | 检查项 | 判定 |
| C1 | task-breakdown 的 ID、名称、依赖与 backlog 完全一致 | 不一致 = 🔴 BLOCKER |
| C2 | 架构模块被至少 1 条用户故事覆盖 | 未覆盖 = 🟡 WARNING |
| C3 | 交付任务覆盖所有架构模块 | 未覆盖 = 🟡 WARNING |
| C4 | 无前后矛盾假设(同一假设在不同文件中状态不同) | 矛盾 = 🟡 WARNING |
| C5 | HOME.md 风险摘要中的风险 ID 与 risk-register.md 一致 | 不一致 = 🟡 WARNING |
| C6 | task-breakdown 表中无重复行(同一任务 ID 多次出现) | 重复 = 🔴 BLOCKER |
| C7 | HOME.md 声明的阶段与文档实际成熟度一致 | 不一致 = 🟡 WARNING |
C7 增强规则:
- 如果
implementation-kit/已存在且有内容,但 HOME.md 阶段写"架构设计中"→ WARNING(应至少为"设计验收") - 如果
wiki/05-delivery/task-breakdown.md有完整任务拆解,但 HOME.md 阶段写"需求澄清中"→ WARNING
⑤ 可追踪性(Traceability)
| ID | 检查项 | 判定 |
| T1 | 每个任务可映射到 prompt(prompts/ 或 implementation-kit/prompts/) | 无映射 = 🟡 WARNING |
| T2 | 每个任务可映射到 acceptance | 无映射 = 🟡 WARNING |
| T3 | 关键决策有记录(HOME.md 决策表 或 wiki/08-history/decision-log.md) | 无记录 = 🟡 WARNING |
| T4 | acceptance-plan.md 中的追踪矩阵完整(任务→prompt→acceptance→report 全链路) | 不完整 = 🟢 INFO |
| T5 | express-incubate 孵化项目存在 execution-log | 缺失 = 🟡 WARNING |
T5 执行记录检查:
- 检查
reports/execution/express-incubate-*/summary.md是否存在 - 如果 implementation-kit 存在但 execution-log 缺失 → 🟡 WARNING(说明孵化被中断或 skill 版本不支持)
- 如果项目使用旧流程(无 express-incubate),检查是否有
reports/execution/目录及 README 说明 - execution-log 是 express-incubate 项目的必需产出,缺失表明孵化流程不完整
⑥ 规范合规(Compliance)
| ID | 检查项 | 判定 |
| L1 | 所有文档使用中文 | 非中文 = 🟡 WARNING |
| L2 | 关键阶段存在对应 reports/stage-review/*.json | 缺失 = 🟢 INFO |
| L3 | implementation-kit 含所有工具适配文件(CLAUDE.md、AGENTS.md、.cursor/rules/、.github/copilot-instructions.md) | 缺失某工具 = 🟢 INFO |
| L4 | user-stories.md 使用标准格式("作为<角色>,我希望<能力>,以便<目标>") | 未使用 = 🟢 INFO |
| L5 | 核心页面遵循统一页面骨架(本页目标、当前结论、依赖与关联、开放问题/风险、下一步) | 缺失槽位 = 🟢 INFO |
| L6 | deployment.md 含最低内容:环境划分 + 依赖服务 + 部署步骤 + 回滚入口 | 缺失 = 🟡 WARNING |
L5 核心页面清单(deep 模式下检查):
wiki/00-overview/vision.mdwiki/00-overview/scope.mdwiki/03-architecture/system-design.mdwiki/04-data-and-api/api-spec.mdwiki/07-operations/deployment.md
L6 最低内容判定:
- 环境划分:至少有 local 和 production/docker 两种环境的说明
- 依赖服务:列出运行时依赖(含版本要求)
- 部署步骤:可执行的命令序列
- 回滚入口:明确的回滚方式
⑦ 阶段合规(Phase)
| ID | 检查项 | 判定 |
| P1 | HOME.md 声明的阶段与实际文档状态匹配(参见 C7) | 不一致 = 🟡 WARNING |
| P2 | 如果声明"设计完成",所有最小文件有实质内容 | 空文件或仅占位 = 🔴 BLOCKER |
| P3 | 已存在的页面有实质内容(非仅骨架/TODO/待补充占位) | 空洞页面较多 = 🟡 WARNING |
| P4 | 无矛盾状态信号(如 "当前目标:实现 PoW 求解器" 但 implementation-kit 已存在) | 矛盾 = 🟡 WARNING |
P3 实质内容判定:
- 页面非空
- 至少有一节具体内容(不只是"待补充"或"TODO")
- 无大面积空白表格
Phase 3: 输出报告
- 汇总每个维度的 PASS/WARNING/BLOCKER/INFO
- 计算总分(每个 WARNING -5,BLOCKER -15)
- 对每个异常给出:检查项 ID、描述、具体文件路径/行号、修复建议
- 按严重性排序:BLOCKER → WARNING → INFO
- 生成
reports/audit/<timestamp>.json和.md - 呈现给用户;如存在需要业务决策的 BLOCKER,再使用当前环境可用的用户输入方式确认
评分计算
总分 = 100
- BLOCKER × 15
- WARNING × 5
- INFO × 0
最低 0 分| 分数 | 判定 |
| 90-100 | ✅ PASS — 可以进入下一阶段 |
| 70-89 | ⚠️ WARNING — 有改进空间,不阻塞 |
| 40-69 | 🔴 NEEDS FIX — 建议修复后再推进 |
| 0-39 | 🚫 BLOCKED — 必须修复 |
检查项密度矩阵
| 深度 | 包含检查项 | 检查项数量 |
| quick | S1-S4, F1-F5, N1, P1-P2 | 12 项 |
| standard | 全部 S/F/N/C/T/L/P | 36 项 |
| deep | standard + L5 逐文件内容检查 | 36+ 项 |
Target Pages
<项目根目录>/reports/audit/<timestamp>.json<项目根目录>/reports/audit/<timestamp>.md
Enriched Behavior
执行前环境检查
执行审计前,必读以下文件以获取基准:
- 仓库根目录
../../../PROJECT_WIKI_TEMPLATE.md— 获取唯一标准结构定义 <项目>/HOME.md— 获取项目当前阶段声明- 根据深度决定是否读取所有 wiki 文件
自检机制
完成审计后,对照本 skill 的检查项清单自查:是否每一项都实际执行了(而非仅凭印象判断)。
Changelog
| 日期 | 变更 | 原因 |
| 2026-07-15 | 删除内置模板副本,改读仓库根模板;评审状态改查报告;修正工具文件名和检查项计数 | 消除模板漂移和审计规则自相矛盾 |
| 2026-04-30 | v2.2 更新 F6 阶段映射表:新增一致性审查、设计并行批次(ui/security/performance/test/threat-scenarios)、量化验收等新阶段 | express-incubate v8 新增阶段,旧映射表已过时 |
| 2026-04-27 | v2.1 新增 T5 检查项:express-incubate 孵化项目必须存在 execution-log | 复盘发现多个项目因 session 中断或 skill bug 导致 execution-log 缺失,影响孵化追溯性 |
| 2026-04-26 | v2.0 重大增强:新增 11 项检查(S4/F5-F7/C5-C7/L4-L6/P3-P4/T4),补充 Input/Output Schema、Minimum/Complete Output、Enriched Behavior | 对 deepseek-web-proxy 审计发现原有 skill 遗漏了重复 ID、面板不完整、阶段标记过期、文档格式合规等 6 类问题 |
| 2026-04-26 | 新建 skill | 全项目级别审计:7维度×3深度×4严重级别,填补 stage-review 无法覆盖的跨文件一致性问题 |
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

