AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Qa Audit

skill-blueprintos-analysis-to-delivery-qa-audit · by BlueprintOS

QA 审计 — 跑全量文档质量检查,按 P0/P1/P2 严重度分级输出审计报告。设计完成、代码交接前调用本 skill,作为进入实施的最后一道硬门控,确保 P0=0 才放行。

No reviews yet
0 installs
34 views
0.0% view→install

Install

$ agentstack add skill-blueprintos-analysis-to-delivery-qa-audit

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-blueprintos-analysis-to-delivery-qa-audit)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
2mo ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Qa Audit? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

QA-Audit — 文档质量审计

Contract

  • 输入: 已完成的 01-09 编号文档,项目配置文件,知识库 / 合规 / 技术栈指针
  • 输出: 09-QA审计报告.md
  • 门控: python3 scripts/full-qa-audit.py 返回 P0=0;整改循环闭环;用户签字
  • Required rules: stage-gate, no-field-guessing, no-self-invent, sql-dialect, doc-numbering, goal-boundary
  • Required paths: knowledge-path, tech-stack-path, compliance-path, doc-naming-path
  • 下一步: /handoff

适用场景

  • dev-design 全部产出已签字
  • 设计回测已通过
  • 进入代码交接前的最后一道闸

流程步骤

1. 跑全量审计

python3 /analysis-to-delivery/scripts/full-qa-audit.py 

2. 审计维度(6 大类)

| # | 维度 | 工具 | |---|---|---| | 1 | 文档格式(frontmatter / H1 / 章节 / 链接 / 占位符) | doc-validate.py | | 2 | 文档编号冲突 | full-qa-audit.py | | 3 | 核心文档完整性(01/05/07 必备) | full-qa-audit.py | | 4 | SQL 方言一致性 | sql-dialect-check.py | | 5 | 字段映射一致性 | field-alignment-check.py | | 6 | 删除/修改的精度 | 人工 + field-alignment-check.py |

3. 分级处理

| 等级 | 含义 | 处理方式 | |---|---|---| | P0 | 致命:导致交接后开发返工 | 必须修复 | | P1 | 重要:影响开发效率 | 建议修复 | | P2 | 提示:锦上添花 | 可选 |

4. 修复迭代

  1. 首次失败 → 读错误输出 → 定位问题文件/行号 → 修复 → 重新跑
  2. 二次失败 → 扩大搜索范围,检查关联文件 → 修复 → 重新跑
  3. 三次失败 → 停止自动修复,输出完整错误报告,等用户决策

5. 写审计报告

输出 09-QA审计报告.md:

# QA 审计报告:{项目名}

> 审计时间:{日期}
> 审计人:{name}

## 审计概览

- 文档总数:{N}
- P0:{N}  ❌ 必须修复
- P1:{N}  ⚠️ 建议修复
- P2:{N}  💡 可选

## 6 大类审计结果

### 1. 文档格式
✅/⚠️/❌

### 2. 文档编号
✅/⚠️/❌

### 3. 核心文档完整性
✅/⚠️/❌

### 4. SQL 方言
✅/⚠️/❌

### 5. 字段映射
✅/⚠️/❌

### 6. 删除/修改精度
✅/⚠️/❌

## P0 问题清单

| 文件 | 行号 | 问题 | 修复方案 |
|---|---|---|---|
| | | | |

## 修复完成确认

- [ ] P0 全部修复
- [ ] 重新跑审计,P0=0
- [ ] 用户签字进入 `/handoff`

调用的 rule

  • rules/stage-gate — 阶段 8 门控
  • rules/sql-dialect — SQL 方言
  • rules/doc-numbering — 文档编号

结束条件

  • [ ] 6 大类审计全部跑过
  • [ ] P0=0
  • [ ] 审计报告签字(用户 + 接收方)

反模式

  • ❌ P0 不修复就交付 — P0 必须=0 才能签字进入交接
  • ❌ 只查 1 类(只查文档/只查代码) — 必须 6 大类(文档/流程/字段/合规/代码/性能)全跑
  • ❌ 审计结果"看起来 OK" — 必须有证据(具体文件 + 行号 + 引用规范)
  • ❌ 不分 P0/P1/P2 — 必须按严重程度分级,否则后续修复无法排期
  • ❌ 跳过代码 vs 设计对比 — 必须检查"代码实现是否偏离设计"
  • ❌ 审计报告无签字栏 — 必须含用户 + 接收方两栏(参考 rules/stage-gate.md 第 8 门)
  • [ ] 进入 /handoff

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.