AGENT SKILL
security-review
暂无简介
- License 许可
- MIT
- 最近更新
- 2026-07-21
- Stars
- ★ 835
- Forks
- 83
02 资源文件 7
- 脚本index.mjs7.7 KB
- 其他pyproject.toml114 B
- 参考文档references/security-checklist.md1.4 KB
- 脚本run.sh7.0 KB
- 脚本sanity.sh937 B
- 参考文档skill.manifest.json2.9 KB
- 参考文档skill.ui.json956 B
03 SKILL.md 正文
Security Review
Runs a bounded static security audit and writes a local threat-model report under .friday.
Design pattern: Reviewer
Review Protocol
- Load the checklist from
references/security-checklist.md. - Scan the target scope (repo, diff, or specified files) against each checklist section.
- For each finding, report:
- Severity (critical / high / medium / low)
- Category (auth, input validation, data protection, dependency, infra)
- Location (file:line or config path)
- Evidence (the vulnerable pattern found)
- Remediation (specific fix with code example)
- Group findings by category, then by severity.
- Write the threat-model report to
.friday/security-review-<timestamp>.md.
Constraints
- Stay within static analysis — do not execute code or make network requests.
- Flag uncertainty explicitly ("potential issue — needs manual verification").
- Prioritize actionable findings over exhaustive coverage.
Typical triggers:
run a security reviewthreat model this repoaudit auth and token safety