Agent Team Kit

**Judgement** Agent Team Kit is a strong process-layer skill for OpenClaw setups that want multi-agent teams to run with minimal human coordination. It doesn’t add complex logic, but it does provide a well-thought-out structure (files, roles, and heartbeat hooks) that many agent projects lack. Best suited for users comfortable letting agents coordinate via shared markdown state. **Strengths** - Clear, opinionated structure for roles, queues, and status tracking. - Heartbeat integration to keep discovery, triage, and execution loops alive without constant human oversight. - Simple to adopt (copy templates + edit ROLES/HEARTBEAT) and easy to customize for different domains. **Risks / Limitations** - Relies on agents correctly adhering to the process; there’s no enforcement mechanism beyond convention and heartbeat checks. - May be overkill for very small or short-lived projects where a full intake/triage loop is unnecessary. - No built-in access control or safety guardrails around what agents can discover, triage, or execute; must be combined with separate safety/permission mechanisms. - If not periodically pruned, OPPORTUNITIES/BACKLOG can accumulate noise and stale items, reducing the effectiveness of self-service queues. **Recommended Scenarios** - Medium to long-running OpenClaw projects where multiple agents work in parallel and you want them to manage their own work stream. - Teams interested in continuous discovery (ideas, experiments, improvements) that flow into a structured backlog without manual task entry by humans. - Projects where you already use HEARTBEAT and want it to monitor team health (stale opportunities, empty READY queue, stuck tasks).