- Home
- Skills
- Agents & Orchestration
- Agent Skill Trigger Doctor Pro: CI gate for vague SKILL.md
Works with the AI Agents you already use
Agent Skill Trigger Doctor Pro: CI gate for vague SKILL.md
Use when a skill library collides, descriptions get dropped under budget pressure, or you need a CI gate that fails vague or colliding SKILL.md descriptions before merge.
$9.99
Agent Skill Trigger Doctor Pro: CI gate for vague SKILL.md
Example session with this skill installed
Scan ~/.agents/skills for trigger collisions and description-budget risk. Fail CI if any description is topic-shaped or >60% token-identical to a sibling.
- Read your context and instructions
- Compiled the agent trigger doctor
- Generated the document
Catalogue scan summary
| Skill | Shape | Soft budget | Collision | Verdict |
|---|---|---|---|---|
| sql-query-reviewer | FAIL topic-shaped | PASS | overlaps db-perf-helper 0.64 | FAIL |
| db-perf-helper | WARN weak WHEN | PASS | overlaps sql-query-reviewer 0.64 | FAIL |
| git-commit-writer | PASS trigger-first | PASS | none >0.60 | PASS |
| (dropped from agent context) | — | OVER hard budget cohort | — | WARN budget |
Collision report (excerpt)
| Pair | Shared tokens | Differentiation ask |
|---|---|---|
| sql-query-reviewer ↔ db-perf-helper | sql, query, performance, optimize | Name concrete artifacts: injection/N+1 vs EXPLAIN/index plans |
CI gate
$ python3 scripts/scan_skill_catalogue.py ~/.agents/skills --fail-on vague,collide
FAIL sql-query-reviewer: R1.3 topic-shaped description
FAIL db-perf-helper: collision>0.60 with sql-query-reviewer
exit 1
Next steps
- Apply
scripts/suggest_trigger_rewrites.pysuggestions one skill at a time - Re-run scan until exit 0
- Wire
templates/github-action-skill-trigger.ymlon PRs that touch**/SKILL.md

1794008511655-szpl.png
PDF · document
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
Free vs Pro
- Free audits one description in chat with inspect + rewrite + lexical preflight; Pro runs
scripts/scan_skill_catalogue.pyacross a library - Free skips sibling collision unless you paste a list; Pro emits a collision matrix and differentiation prompts
- Free has no budget-drop view across many skills; Pro checks soft/hard description budgets and flags catalogue pressure
- Free has no merge gate; Pro ships a CI template that fails vague or colliding descriptions
- Free leaves rewrites in chat; Pro adds
scripts/suggest_trigger_rewrites.py+ JSON/Markdown export for PRs
Free version: https://www.agensi.io/skills/agent-skill-trigger-doctor-skill-md-that-never-fires
Upgrade Path
Upgrade when Free fixed one skill and the next three still collide — or when reviewers need an artifact, not a transcript.
Free stays useful for a single stuck skill. Pro is for authors and teams who keep a real catalogue.
The Problem
A good one-skill rewrite dies the week you add twelve more tools. Descriptions blur. Two reviewer skills fight for the same ask. Some harnesses quietly drop metadata when the catalogue gets large. A chat checklist cannot fail a pull request. Reviewers inherit the same topic-shaped first line in the next merge.
What You Get
- Catalogue scan (
scripts/scan_skill_catalogue.py): walk a skills directory for shape, budgets, and collisions, with JSON and an exit code for CI - Rewrite suggestions (
scripts/suggest_trigger_rewrites.py): batch WHEN-first description candidates. Bodies stay preserved - Phrase preflight (
scripts/eval_trigger_phrases.py): should-trigger / should-not-trigger suite with Markdown or JSON export. Lexical proxy only, not live harness measurement - Budget check (
scripts/check_description_budget.py): soft (~250) / hard (~512) token estimates and a catalogue-pressure summary - Merge gate (
templates/github-action-skill-trigger.yml): fail the pull request on vague or colliding descriptions - Starter phrase suite plus references for WHEN-first patterns and how to split overlapping siblings by task and artifact
Why This Beats Prompting It Yourself
Asking an agent to "make descriptions better" returns prose. Pro returns exit status. The demo above fails CI on a topic-shaped SQL reviewer that collides with a sibling — the same break a chat session forgets by Thursday.
Use Cases
- Pre-publish catalogue hygiene before a marketplace drop
- Team skill libraries where two reviewers steal each other's triggers
- CI on monorepos that version many
SKILL.mdfiles - After a bulk AI-generated skill pass (topic-shaped descriptions at scale)
Known Limitations
- Lexical / heuristic scoring — not a substitute for live harness measurement
- Does not rewrite skill bodies, scripts, or hooks
- Does not detect settings overrides, wrong install paths, or plugin load failures by itself (scan can warn if paths look empty)
- Collision threshold defaults are tunable; false positives happen on intentionally related skills — review before mass-rewrite
- No sales, ranking, or fire-rate guarantees
- Money-back follows Agensi platform policy; this skill does not promise installs or revenue
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 11 days ago
- Software
Python