- Published on
Skill Installed but Not Triggered? Full Fix Checklist
Install success does not guarantee runtime selection.
If a skill never triggers, use this checklist in order.
1) Confirm canonical file path
test -f ~/.agents/skills/<skill-name>/SKILL.md
ls -la ~/.agents/skills | rg "<skill-name>"
2) Check description clarity
Weak descriptions cause low trigger confidence.
Fix pattern:
- name clear task
- include trigger phrases
- define "when to use / when not to use"
3) Test direct invocation
Prompt:
Use
<skill-name>for this task. Explain why the skill matches and run a non-destructive first step.
If this works, your issue is routing confidence, not installation.
4) Remove overlap temporarily
Disable similar skills and retry the same prompt.
If trigger starts working, adjust scope boundaries across overlapping skills.
5) Verify runtime permissions
Even when selected, skills can fail and appear as "not used."
Check:
- permission policy
- blocked commands
- missing dependencies
6) Lock final fix
Before re-enabling full stack:
- run 3 successful test prompts
- confirm consistent selection behavior
- document trigger cues for the team
Related:
Written by OpenClaw Community Editorial Team. Last reviewed on . Standards: Editorial Policy and Corrections Policy.