a2a-security-review
Review agent-to-agent (A2A) / multi-agent-system trust: peer identity and authentication, message integrity, capability-negotiation trust, and delegation-chain…
Analyze a project's dependencies (software composition analysis): generate/inspect an SBOM, find known-vulnerable and risky components, and prioritize upgrades. Use when reviewing third-party/open-source risk in a codebase or build.
$ npx -y skills add jassics/awesome-claude-security --skill sca-review --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/sca-reviewContext preview
The summary Claude sees to decide when to auto-load this skill.
Analyze a project's dependencies (software composition analysis): generate/inspect an SBOM, find known-vulnerable and risky components, and prioritize upgrades. Use when reviewing third-party/open-source risk in a codebase or build.
name: sca-review description: >- Analyze a project's dependencies (software composition analysis): generate/inspect an SBOM, find known-vulnerable and risky components, and prioritize upgrades. Use when reviewing third-party/open-source risk in a codebase or build.
A prioritized dependency-risk view: which components are vulnerable or risky, how reachable/exploitable they are, and the upgrade path — backed by an SBOM.
1. **Inventory** — generate or ingest an **SBOM** (e.g. CycloneDX/SPDX) covering direct and transitive dependencies; capture versions and licenses. 2. **Match vulnerabilities** — map components to known advisories (CVE/GHSA/OSV). Record severity and whether a fixed version exists. 3. **Assess reachability/exploitability** — is the vulnerable function actually used? Prefer reachability over raw CVE counts to cut noise. Note runtime vs. build-only and dev-only dependencies. 4. **Check hygiene** — unmaintained/abandoned packages, suspicious or typosquatted names, integrity/provenance (lockfiles, signing), and risky licenses. 5. **Prioritize remediation** — reachable + high-severity + fix-available first; plan upgrades and watch for breaking changes.
A prioritized table: component · version · advisory · severity · reachable? · fixed-version · action, plus the SBOM and a short upgrade plan. Confirmed issues → `security-reporting:finding` (maps to OWASP A06 / API-adjacent supply-chain risk).
Rank by reachability and exploitability, not CVE count — most flagged CVEs aren't reachable. Don't ignore transitive deps (where most real risk hides) or unmaintained packages (a risk even without a current CVE). For deeper provenance/SLSA/signing, see `supply-chain-security` (roadmap).
A Claude Code plugin marketplace for the full cybersecurity & GenAI-security lifecycle — from recon and threat modeling to detection engineering, GRC, and CISO-level strategy. A pentester knows which OWASP test bends a broken-access-control endpoint.
Repo: jassics/awesome-claude-security
Review agent-to-agent (A2A) / multi-agent-system trust: peer identity and authentication, message integrity, capability-negotiation trust, and delegation-chain…
Test the agent execution harness/runtime itself — LangChain/LangGraph, AutoGen, CrewAI, custom ReAct-style loops, or computer-use/browser-use agents — for…
Assess an autonomous / tool-using AI agent for security end-to-end: tool privileges, autonomy and approval boundaries, excessive agency, memory/state…
Test what an AI agent will actually do without human confirmation, including under injected-goal / prompt-injection scenarios, to validate its autonomy and…
Review the security of MCP (Model Context Protocol) servers/clients an agent uses: server trust tier, tool/resource description and result poisoning,…
Inventory the tools/functions an AI agent can call and audit their privileges, side effects, and approval requirements to find excessive-agency and…