| name | t3mt-remediation-ops |
| description | Use this skill to analyze recent T3MT/T3FAP failures, propose recovery actions, and execute focused remediations for unhealthy plugins, failed tasks, and blocked automations. |
T3MT Remediation Ops
Use this skill when the user wants the assistant to diagnose failures and move toward recovery instead of only reporting monitor data.
Follow t3mt-sidecar-automation for mutation, confirmation, audit, and secret-handling rules.
Use this skill for
- Turning recent monitor failures into concrete recovery candidates.
- Re-enabling one unhealthy plugin after current state inspection.
- Re-running or recovering one task or execution with a clear target.
Do not use this skill when
- The user only wants a read-only status snapshot. Use
t3mt-monitor-ops.
- The user already knows the exact plugin, task, or resource mutation they want. Use the matching domain skill directly.
- The request is a broad multi-object cleanup or destructive reset. Escalate and confirm first.
Common commands
python scripts/t3mt-remediation-ops.py analyze limit=20
python scripts/t3mt-remediation-ops.py plugin-recover plugin_id=task.transfer
python scripts/t3mt-remediation-ops.py task-recover task_id=task-demo auto_run=true
python scripts/t3mt-remediation-ops.py execution-recover execution_id=exec-demo auto_run=true
What it does
- Reads plugin health and recent executions.
- Produces prioritized recovery candidates with suggested commands.
- Can re-enable a plugin, re-run a task, or recover from a failed execution.
- Keeps recovery focused on one plugin or one task at a time.
Guidance
- Use
analyze first when the failure source is still unclear.
- Prefer
plugin-recover for disabled or unhealthy plugins.
- Prefer
execution-recover when you have a specific failed execution and want the assistant to trace it back to the owning task.
- In
full-access, single-target recovery actions may run automatically after the current state is read.
