← ClaudeAtlas

red-teamlisted

Task the reviewer with actively breaking a system, API, validation layer, prompt pipeline, policy boundary, or logic flow that the active primary (or the user) has just built. The verifier generates concrete adversarial inputs — exact strings, payloads, scenarios — designed to bypass controls, crash the system, or trigger misbehavior. Use when the user says "red-team this," "try to break this," "attack this," "stress-test this," "adversarial test this," "find ways this could fail," "what could go wrong with this validation," "break my parser," "break my prompt," "find bypasses," or similar break-this-system framings. Also offer this proactively when the active primary has just shipped or is about to ship a security boundary, input validation, authentication flow, content-moderation policy, prompt pipeline, rate-limiting rule, payment validator, or any control surface where the cost of an undiscovered bypass is high.
sumitake/agent-collab · ★ 0 · AI & Automation · score 63
Install: claude install-skill sumitake/agent-collab
## Unified runtime invocation Resolve the **plugin root** from this loaded file: `SKILL.md` is at `<plugin-root>/skills/<skill-name>/SKILL.md`. Invoke only `python3 "<plugin-root>/coordinator.py"` and send one bounded JSON routing request on EOF-delimited stdin, without a PTY. Use the Python invocation example in the **Routing request** section in `<plugin-root>/README.md` and the co-packaged manifest's signed `wire_contract`; never invent fields or provider actions. Supply one caller-defined work unit per independently useful deliverable, with this skill's logical action and a bounded opaque payload. Use `depends_on` only for actual dependencies. Honor an operator-named provider with `explicit_target`. For an authorized independent review or governance task without an operator-named provider, also use that field to bind the caller-verified distinct reviewer selected by the caller or designated by the workflow. Carry the same target into planning and live dispatch; verify returned native lineage before accepting independence. Otherwise use normal untargeted routing. Choose quality and effort for the workload; include context/output token estimates when known. Read the current manifest digest and actual cwd device/inode; do not copy example values. The runtime owns its timeout; do not wrap it in a shorter fixed timeout. Repository identity, source-head verification, disposable copies, patch capture, and cleanup remain caller-owned where applicable. The shim runs standalone fr