AGENT SKILL

security-audit

Use for source-grounded security questions, vulnerability investigation, focused security reviews, or explicit codebase audits and pen tests. Not dependency scanning or mandatory general review.

License 许可
MIT
最近更新
2026-11-12
Stars
★ 4
Forks
1

01 前言 Front Matter

name
security-audit
description
Use for source-grounded security questions, vulnerability investigation, focused security reviews, or explicit codebase audits and pen tests. Not dependency scanning or mandatory general review.

02 资源文件 1

03 SKILL.md 正文

Security audit

Apply these host rules, then read the complete upstream skill.

  • Resolve the skill directory to the upstream file's parent, not this adapter or the Codex proxy. Resolve companion guides, schema, and validators from there.
  • Default to guidance for focused questions/reviews. Explicit codebase audits, pen tests, comprehensive reviews, or report requests select the six-phase workflow and retained artifacts.
  • Delegation needs explicit user consent; an audit request alone is not consent. Without consent or delegation tools, provide source-only guidance inline and disclose partial coverage, not a full audit or independent verification. Obtain consent before a full run. Preserve host model selection and budgets.
  • Keep upstream sandbox controls, local-only scope, dummy data, write isolation, evidence gates, and validators. Missing execution controls mean source inspection plus a needs-validation blocker, not unsandboxed target execution or live probes. An audit describes fixes; implement only if requested.

Provenance and refresh. Installation enables no audit hook, background work, or automatic security-review lane.

04 同仓库更多 Skills