Use when the user asks for a deep, exhaustive, multi-pass, or variance-reducing repository-wide or scoped-path Codex Security scan. Run repeated complete independent Standard scans with the Codex Security deep-scan tool, which aggregates their validated findings and prepares the canonical artifacts; then complete the same scan once. Do not use for PRs, commits, branch diffs, or working-tree diffs.
Use when hardening APIs: HTTPS, HMAC 签名, timestamp+nonce 防重放, CORS, 脱敏, JWT none 算法, 密钥不进 URL. For SQL injection and resource 越权 WHERE, use backend-safe-check.
Security scan workflow — dependency audit, OWASP checklist, secrets scan, vulnerability report. Applies software-engineer role with security focus. Use standalone or as part of code review.
The Atelier Security Law — OWASP-grade defense for every request path (injection, XSS, CSRF, tenant isolation, rate limiting, transport). Load before writing ANY endpoint, form, or data-processing code — not just auth. Auth token/session mechanics live in [[auth-law]]; this is everything else.
Mobile app security for Kotlin Multiplatform beyond Android-only R8 obfuscation — certificate/SSL pinning (expect/actual, no cross-platform Ktor support exists natively), root/jailbreak/tamper detection via freeRASP's real KMP variant, encrypted local storage via KSafe, iOS/ Kotlin-Native release-binary symbol stripping, and an OWASP Mobile Top 10 2024 coverage map across this collection. Does NOT cover Android-specific R8/ProGuard obfuscation — that's kmp-proguard-r8's own scope, cross- referenced here rather than duplicated. Does NOT cover secrets-in-source scanning (gitleaks pre-commit) — that's kmp-setup-hooks Option F.
Focused API security auditor for REST, GraphQL, and webhook endpoints in any stack. Use when the user is building or reviewing an API layer, mentions rate limiting, API keys, webhooks, CORS, or specifically wants "API security" checked — as opposed to a full-project audit.
Deep-dive security audit of a technical architecture diagram. Focused exclusively on authentication, authorization, credential protection, network boundaries, and data classification. Does NOT score overall quality — produces a prioritized security finding list. Use after arch-validate when you want a security specialist's deep cut.
Read-only security audit of code for SQL injection, XSS, auth/authz flaws, input validation gaps, sensitive data exposure, and insecure cryptography. Surfaces findings without modifying code.
When the user needs a security assessment — threat modeling, vulnerability review, auth flow audit, dependency scanning, or says "is this secure", "review for vulnerabilities", "threat model", "security audit", "pen test prep".
Use when the user asks for a deep, exhaustive, multi-pass, or variance-reducing repository-wide or scoped-path Codex Security scan. Run repeated complete independent Standard scans with the Codex Security deep-scan tool, which aggregates their validated findings and prepares the canonical artifacts; then complete the same scan once. Do not use for PRs, commits, branch diffs, or working-tree diffs.
Secure an agent system as a permission surface — what authority each tool call runs under, where the confused deputy sits, which controls are enforced outside the model and which are only prompt text, and how far one bad call reaches. Use when granting an agent tools or credentials, wiring in MCP servers or subagents, reviewing an agent that acted beyond what the requester could have done, or before letting an agent touch a shared or production system. Not for wording the ingestion trust boundary in detail (prompt-injection-defense), not for scoping what an agent is for (agent-design), and it never grants an agent permission it did not already have.
Plain text files in a repository that tell a coding agent how the project works: commands to run, conventions to follow and things to avoid. CLAUDE.md, AGENTS.md, cursor rules and skills are the common kinds.
CLAUDE.md or AGENTS.md?
CLAUDE.md is read by Claude Code. AGENTS.md is an open format that Codex, Cursor and other agents read. Many projects keep one and point the other at it.
What is a skill?
A folder with a SKILL.md that describes one capability, such as filling PDFs or reviewing code. The agent loads it only when the task calls for it.
Can I search my own team's files too?
Your agents already can, over MCP, limited to the files you're allowed to read. Searching them from this page is coming.
OpenAISkill
