red-team-assessmentlisted

# Red Team Assessment Submit code, designs, or configurations to one or more LLMs for adversarial security analysis. Give reviewers file/tool access appropriate to the CLI and request `sqry`, `exa`, and `ref_tools` when code search, CVE research, or documentation checks are needed. ## Dispatch Defaults Apply these on every dispatch unless the caller has explicitly overridden a rule in the current turn: 1. **Omit `model`** — let the gateway use its configured default per CLI. Nominating a model risks deprecated IDs (`o3`, `o3-pro`, `gpt-4o`, …) and capability mismatches. Only nominate when the caller has explicitly named a specific variant. 2. **`approvalStrategy:"mcp_managed"`** is the skill dispatch default (the gateway schema default is `"legacy"`). It runs the gateway gate first; Claude then uses `bypassPermissions`, Gemini uses `yolo`, and Codex still needs `fullAuto:true` for autonomous file/shell work. Add `mcpServers:["sqry","exa","ref_tools"]` when research tools are needed. 3. **No wallclock timeout; poll every 60 s** — red team assessments are thorough and routinely run for 5–20 minutes. Do **not** cancel for "taking too long." `idleTimeoutMs` (no-output safeguard) is separate. 4. **Iterate until unconditional APPROVED** (review dispatches only) — every red team prompt must end with "End with PASS (no critical/high findings) or FAIL with findings" (the PASS/FAIL verdict is the red-team equivalent of APPROVED/NOT APPROVED; treat PASS as APPROVED). On FAIL, run th