🔵 Stable🔷 SkillSpec L3⚙ ships an executable helperpm-skill-promoter
Generate an evals file for a drafted skill and measure whether its trigger description fires on the right requests: five to eight phrases that should trigger it, five that should not, three golden input and output pairs, then precision and recall against the description, with a rewrite loop when precision is below 0.8. Use when asked to test this skill's triggers, write evals for the new skill, check the description fires correctly, or as the third step of the promote loop. Produces an evals.json beside the skill, a case in the library's evals, and a precision and recall report.
“Test the triggers on the new skill.”“Write evals for release-notes-from-git-log.”“Does this description fire on the right prompts?”“Run the third step of the promote loop.”
What to give it
▸The drafted skill — path to its SKILL.md
▸The scan pattern it came from — , for the real phrasings; otherwise ask the user for five ways they would say it
▸Two or three near-miss jobs — things that sound similar but should go to a different skill (these become the hardest negatives)
✅ The bar it holds itself to
Every skill in this library self-verifies — these are this skill's own quality checks, straight from its definition.
✓5 to 8 positives, 5 negatives with at least two near misses, 3 goldens
✓Precision 0.8 or above on the final run, recall reported honestly
✓Every rewrite is shown, not just the winning description
✓The evals/cases.json entry uses the first golden's input
✓Nothing in the phrases names a real person, company or dataset
⚠️ What it refuses to do
**Easy negatives only.** If nothing in the negatives shares a word with the skill, the test proves nothing.
**Chasing recall by stuffing the description.** A description that lists forty phrases fires on everything.
**Skipping the library case.** The coverage ratchet fails the pull request without it.
**Testing against the body instead of the description.** Only the description is read at trigger time.
Install
npx pm-claude-skills add --agent claude # or codex · cursor · gemini · hermes
# or one-line MCP (every skill, any client):
claude mcp add pm-skills -- npx -y pm-claude-skills-mcp