Complements the validation fix from c6fe9d71 - the parseAgentIDFields
function now also scans right-to-left for known role tokens instead
of using fixed position parsing.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Previously, `bd agent state <agent> <state>` would fail if the agent bead
didn't exist in the database. This caused issues when `gt sling` tried to
update agent state for newly spawned polecats.
Now when the agent doesn't exist:
1. Parse role_type and rig from the agent ID (e.g., gt-gastown-polecat-nux)
2. Auto-create the agent bead with type=agent
3. Add role_type and rig labels for filtering (bd list --label=role_type:polecat)
4. Continue with the state update
This enables:
- Work history accumulation per polecat name
- Skill/success tracking over time
- `bd list --type=agent` to see all agents
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add agent commands for self-reporting state:
- bd agent state <agent> <state>: Update agent state and last_activity
- bd agent heartbeat <agent>: Update last_activity timestamp only
- bd agent show <agent>: Display agent bead details
States: idle, spawning, running, working, stuck, done, stopped, dead
Also adds AgentState and LastActivity fields to UpdateArgs in RPC protocol.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
