It reports; it never fixes the audited project and never changes the user's settings. Every finding carries its evidence (file:line), passes a read-only refutation pass, and cites the standard entry it maps to, as read on the source named in the references.
Agentic Security Skill
Overview
Project configuration for an agent is executable: a .claude/settings.json hook, an .mcp.json
server command or an env override runs on the developer's machine. Check Point Research
showed this end to end (CVE-2025-59536, CVE-2026-21852). This skill turns the vendor
documentation into checklists an auditor can run read-only.
Audit Areas
| Area | What is checked | Reference |
|---|---|---|
| Trust boundary | Hooks, env, helpers, MCP auto-approval and base-URL overrides that act before trust |
trust-boundary.md |
| Hook supply chain | Plugin hooks.json, plugin.json hooks / mcpServers / userConfig |
hooks-plugins.md |
| MCP trust | Every server source, publisher, pinning, env and tokens, tool descriptions | mcp-trust.md |
| Prompt injection | Untrusted content reaching the agent, instruction files that obey observed content | prompt-injection.md |
| Permissions | Broad allow rules, deny-rule limits, bypass and auto modes, saved approvals | permissions.md |
| Secrets | Keys and tokens in settings, .mcp.json, hooks, memory and lesson files, .env |
secrets.md |
| CI agents | AI review or agent actions running on untrusted pull requests | ci-agents.md |
Workflow
- Inventory the agent surface of the target path:
.claude/,.mcp.json,.claude-plugin/,hooks/,skills/,agents/,CLAUDE.md, rule and memory files,.github/workflows/, and the equivalent folders of other agent tools present in the repository. - Run the checklist of each area in scope (audit-checklist.md).
- Score each candidate: severity, confidence 1-10, CWE and standard entry from standards-map.md.
- Verify: run the refutation pass in verification.md on every candidate. Confidence 8 or more becomes a finding; lower goes to "needs manual review".
- Report with audit-report.md; the caller writes it to
docs/security/YYYY-MM-DD-<scope>.mdinside the audited repository.
Critical Rules
| Rule | Why |
|---|---|
| Report only, never fix or change settings | The audit must not alter the trust state it measures |
| Never print a secret value | Report file:line and the secret type only |
Read ~/.claude, ~/.codex and similar only when the owner asks, and read-only |
User-level config is outside the audited repository |
| Cite every fact from the reference files | Vendor behavior changes by version; no claim from memory |
| A static read never proves exploitability | See verification.md |
References
- Trust boundary — what a repository can run before trust, the two CVEs
- Hooks and plugins — plugin hook supply chain
- MCP trust — server inventory and trust checks
- Prompt injection — injection paths into the agent
- Permissions — rules, modes and their documented limits
- Secrets — secrets in agent configuration and memory
- CI agents — agents in GitHub Actions
- Standards map — OWASP LLM / Agentic entries and CWE, as read
- Verification and limits — refutation pass, what the skill does not prove
- Audit checklist · Audit report