a2a-security-review
Review agent-to-agent (A2A) / multi-agent-system trust: peer identity and authentication, message integrity, capability-negotiation trust, and delegation-chain…
Plan and run a purple-team exercise: collaboratively emulate specific ATT&CK techniques and measure whether detection and response actually work, then close the gaps. Use to validate defensive coverage against real adversary behavior. Authorized environments only.
$ npx -y skills add jassics/awesome-claude-security --skill purple-team-exercise --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/purple-team-exerciseContext preview
The summary Claude sees to decide when to auto-load this skill.
Plan and run a purple-team exercise: collaboratively emulate specific ATT&CK techniques and measure whether detection and response actually work, then close the gaps. Use to validate defensive coverage against real adversary behavior. Authorized environments only.
name: purple-team-exercise description: >- Plan and run a purple-team exercise: collaboratively emulate specific ATT&CK techniques and measure whether detection and response actually work, then close the gaps. Use to validate defensive coverage against real adversary behavior. Authorized environments only.
Measured evidence of which adversary techniques your defenses detect and respond to — and a prioritized set of fixes — produced collaboratively (red emulates, blue observes) rather than as a pass/fail contest.
1. **Scope & objectives** — pick techniques to test, driven by threat relevance (`threat-intelligence:threat-actor-profiling`) and known coverage gaps (`detection-engineering:detection-coverage-review`). Define success criteria. 2. **Plan emulation** — map each technique to a concrete, safe emulation (e.g. Atomic Red Team / a known procedure). Confirm authorization, scope, and a rollback plan; coordinate timing with the defenders. 3. **Execute collaboratively** — run each technique; record for each: was telemetry generated? did a detection fire? did an alert reach an analyst? was response timely and correct? Capture timestamps. 4. **Score** — per technique: Detected / Partially / Missed, plus prevention and response outcomes. Identify whether gaps are data-source, detection, or process. 5. **Remediate** — turn misses into detections (`detection-engineering:detection-rule-development`), data-source onboarding, or runbook fixes; re-test to confirm closure.
A results matrix: technique (ATT&CK) · emulation · telemetry? · detected? · responded? · gap type · fix · retest status. Visualize as an ATT&CK heatmap (`security-diagramming`) and report with `security-reporting`.
Purple teaming is collaborative measurement, not a competition — the win is closed gaps, not "red won." Keep emulations safe, authorized, and reversible. Re-test after fixes; an untested fix isn't a closed gap.
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…