| name | agent-retrospective |
| description | Evaluate agent performance after a sprint, review, incident, build cycle, or multi-agent run. Use when the user asks what worked, what failed, which agents were effective, where delegation broke down, or how to improve future agent workflows. |
Agent Retrospective
Turn agent workflow experience into measurable improvements for skills, prompts, policies, and review gates.
Core procedure
- Summarize the mission, agents involved, outputs produced, and validation status.
- Assess each agent against expected output, scope discipline, evidence quality, speed, safety, and handoff quality.
- Identify failure modes: unclear scope, duplicate work, authority confusion, weak tests, stale context, missing evidence, tool misuse.
- Separate systemic issues from one-off mistakes.
- Create improvement actions for skills, policies, templates, hooks, and delegation rules.
- Assign owners and priority for each improvement.
Output format
Produce a concise, auditable markdown artifact. Include:
## Agent Retrospective
**Scope:** ...
**Decision / recommendation:** ...
**Risk level:** low / medium / high / critical
### Findings
- ...
### Plan / matrix / record
| Item | Owner | Scope | Evidence | Gate / next action |
|---|---|---|---|---|
### Escalations
- ...
### Next actions
1. ...
Expected output focus: retrospective with scores, findings, root causes, improvement actions, owner, priority, next validation.
Guardrails
- Prefer the narrowest capable agent and least authority required.
- Do not let an agent approve its own high-risk work.
- Mark assumptions separately from confirmed facts.
- Require human approval for destructive, credentialed, production, external, legal, financial, or security-sensitive actions.
- Do not invent agent capabilities; use the roster, plugin manifests, source files, and explicit user instructions.
Related references
references/agent-retrospective-template.md
references/risk-tier-model.md
references/decision-rules.md