skill-trigger-test
/skill-trigger-test
Tests whether skill descriptions would make the right skill load, against saved requests or the words you really type. Use when someone says "my skills never load", "test my skill descriptions", "why didn't my skill start", "which skill would fire", or "score my skills".
- Author
- hgbeigi ↗
- Repository
- claude-tools ↗
- Licence
- MIT
- Last repository change
- 05.11.2026
Assessment and security
- Files in this package
- 2
- Scripts
- 1
- HTML comments
- 0
- URLs in SKILL.md
- 0
No signals detected. Repository changed on 05.11.2026, without references to retired models or APIs.
MIT, open: use and modify while retaining author attribution.
- 1 scripts in this package. Review them before installing, as Claude may execute them.
Scripts: scripts/trigger_test.py
What it does and when to use it
Tests whether skill descriptions would make the right skill load, against saved requests or the words you really type. Use when someone says "my skills never load", "test my skill descriptions", "why didn't my skill start", "which skill would fire", or "score my skills".
Triggers from the description
Original author description
Tests whether skill descriptions would make the right skill load, against saved requests or the words you really type. Use when someone says "my skills never load", "test my skill descriptions", "why didn't my skill start", "which skill would fire", or "score my skills".
How it works
- Steps
- Limits
SKILL.md preview
---
name: skill-trigger-test
description: Tests whether skill descriptions would make the right skill load, against saved requests or the words you really type. Use when someone says "my skills never load", "test my skill descriptions", "why didn't my skill start", "which skill would fire", or "score my skills".
---
# Skill trigger test
Claude picks a skill by matching a request against its description. This test shows which requests would miss, and which quoted phrases the user never types.
## Steps
1. Find the skills folder. Usually `.claude/skills` in the project, or the one in the home Claude folder. Ask only if neither exists.
2. Choose the measure.
- **Saved requests.** Write 10 to 20 real requests into a JSON file, in the user's own words, each with the skill that should start. Take them from the user, not from the descriptions, or the test proves nothing. Run:
`python "${CLAUDE_PLUGIN_ROOT}/skills/skill-trigger-test/scripts/trigger_test.py" --skills <folder> --requests <file>`
- **The user's own words.** This reads the user's transcripts on this computer, so say so first. Run:
`python "${CLAUDE_PLUGIN_ROOT}/skills/skill-trigger-test/scripts/trigger_test.py" --skills <folder> --transcripts --days 90`
3. Report the score, each miss, and each skill with no quoted phrases.
4. For a miss, propose one change to the description: a quoted phrase in the words the user really typed. Change a description only when the user agrees.
5. Run the same test again and report before and after.
## Limits
Say this with the score: the test reads only the quoted phrases, while Claude reads the whole description, so it is a stand-in for comparing before and after. A skill can still load without a phrase match.
Do not quote transcript text beyond the phrase counts unless the user asks.
Files in this package (2)
- SKILL.md 2 KB
- scripts/trigger_test.py 6 KB
Versions
The database has no other skills named skill-trigger-test.
Database snapshot: 13.11.2026 · package version 098b833e5841 (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 100test-skill-activation
Measure whether agent skills (SKILL.md) actually activate in Claude Code, Codex and OpenCode with doesitfire, then improve the descriptions of skills that do not fire or fire when they should not. Use when the user asks if a skill triggers, why a skill is ignored, to test or… - 100skill-creator
Create or review agent skills: SKILL.md, frontmatter, references, trigger descriptions, overlaps. Use for "make a skill" or a skill that never triggers. - 100skillcheck
Measure whether the user's Claude Code or Codex skills actually trigger in their real setup, next to all their other installed skills and plugins. Use when the user says a skill "doesn't fire", "isn't being used", "gets ignored", "the wrong skill runs", asks to test, audit or… - 100creating-skills
Skill authoring and optimization. Use when creating new skills, editing existing skills, verifying skill quality, or optimizing trigger descriptions. - 84audit-skill-visibility
When the built-in skill-doctor command resolves in this session, prefer it for cost; this skill for unseen ones. Use when: 'why do I never use most of my skills', 'why does Claude never suggest this skill', 'are my skill descriptions being dropped', 'is my skill listing over… - 98skill-hygiene-audit
Audit AI agent skill installations — broken SKILL.md files, duplicate skills shadowing each other across directories, missing or over-long descriptions — and clean them up. Use whenever the user installs or manages skills ("npx skills add", ".claude/skills", ".codex/skills",…