agent-readinesslisted

# Agent readiness - role promotion review Decide whether an agent is ready for a specific role. The unit of review is not "is this model good"; it is "is this agent-runtime-role combination ready for this responsibility." ## Readiness dimensions - **Role fit:** primary author, reviewer, worker delegate, tiebreaker, release agent, headless runner, or specialist. - **Capability evidence:** successful tasks, failed tasks, live smoke tests, benchmarks, and known weak spots. - **Runtime health:** CLI availability, auth state, version currency, quotas, cost model, non-interactive behavior, and timeout handling. - **Tool boundary:** filesystem, network, browser, MCP, shell, credentials, write access, and sandbox behavior. - **Governance:** review independence, merge/tag authority, compliance trace obligations, escalation rules, and CODEOWNERS/operator gates. - **Reliability:** retries, failure signatures, audit logs, resumability, handoff quality, and fallback agent. - **Cost posture:** when to use a cheaper model, when the capability justifies cost, and when to avoid fan-out. ## Workflow 1. Name the exact role and scope being proposed. 2. Collect current evidence from local docs, runtime checks, CI, prior PRs, or recent handoffs. 3. Compare capability needs against known agent strengths and constraints. 4. Identify the minimum pilot or smoke test needed if evidence is incomplete. 5. Decide whether the agent can hold the role now, hold it with conditions, or should stay out of