Skip to content

feat: improve skill scores for 5 core MOOLLM skills#179

Open
rohan-tessl wants to merge 2 commits intoSimHacker:mainfrom
rohan-tessl:improve/skill-review-optimization
Open

feat: improve skill scores for 5 core MOOLLM skills#179
rohan-tessl wants to merge 2 commits intoSimHacker:mainfrom
rohan-tessl:improve/skill-review-optimization

Conversation

@rohan-tessl
Copy link
Copy Markdown

@rohan-tessl rohan-tessl commented Mar 26, 2026

Hey @SimHacker 👋

I ran your skills through tessl skill review at work and found some targeted improvements.

image

Here's the full before/after:

Skill Before After Change
adventure 0% 93% +93%
cat 0% 79% +79%
character 0% 79% +79%
room 0% 74% +74%
abstract 0% 63% +63%

All five skills were scoring 0% due to frontmatter validation errors — the allowed-tools field was formatted as a YAML array instead of a comma-separated string, and abstract had no frontmatter at all. This PR is intentionally capped at five skills to keep it reviewable; the included workflow (below) will surface Tessl feedback on future SKILL.md changes across the rest of your 117-skill library.

Changes summary

Frontmatter fixes (all 5 skills):

  • Converted allowed-tools from YAML array to comma-separated string (the format tessl skill review expects)
  • Moved non-standard frontmatter keys (tier, protocol, lineage, inherits, related, tags, templates, origin) into a metadata block to clear unknown-key warnings

Description improvements (all 5 skills):

  • Expanded descriptions with specific trigger terms and explicit "Use when..." clauses
  • Descriptions now articulate concrete capabilities rather than just naming the skill

Skill-specific changes:

  • abstract: Added full frontmatter (had none — went from no --- delimiters to proper name + description)
  • room: Trimmed from 807 → 499 lines by condensing verbose vehicle, logistics, and data-flow sections into brief summaries with links to related skills. Core room anatomy, regions, and ROOM.yml structure preserved in full

Automated Skill Review Workflow

This PR also adds .github/workflows/skill-review.yml — a GitHub Action that runs on PRs touching SKILL.md files:

  • What runs: tesslio/skill-review detects changed SKILL.md files, runs tessl skill review on each, and posts (or updates) one PR comment with scores and feedback
  • Zero extra accounts: contributors don't need a Tessl login — only the repo's default GITHUB_TOKEN is used to post the comment
  • Non-blocking by default: the check is feedback-only with no surprise red CI, unless you add fail-threshold
  • Not a build replacement: this is review automation for skill markdown only — it doesn't touch your existing pipelines
  • Optional gate: add with: fail-threshold: 70 later if you want PRs to fail on low skill scores
  • Why only five skills here: this PR stays small and reviewable; after merge, every PR that touches SKILL.md gets automatic review comments, so the rest of the library improves incrementally

Honest disclosure — I work at @tesslio where we build tooling around skills like these. Not a pitch - just saw room for improvement and wanted to contribute.

Want to self-improve your skills? Just point your agent (Claude Code, Codex, etc.) at this Tessl guide and ask it to optimize your skill. Ping me - @rohan-tessl - if you hit any snags.

Thanks in advance 🙏

Fix frontmatter validation errors across adventure, cat, character,
room, and abstract skills. All were scoring 0% due to allowed-tools
array format and missing/invalid frontmatter.

Changes:
- Convert allowed-tools from YAML array to comma-separated string
- Move non-standard frontmatter keys (tier, protocol, lineage, etc.)
  into metadata block
- Add frontmatter to abstract skill (had none)
- Trim room skill from 807 to 499 lines (was triggering length warning)
- Expand descriptions with trigger terms and "Use when..." clauses

| Skill     | Before | After | Change |
|-----------|--------|-------|--------|
| adventure |     0% |   93% |   +93% |
| cat       |     0% |   79% |   +79% |
| character |     0% |   79% |   +79% |
| room      |     0% |   74% |   +74% |
| abstract  |     0% |   63% |   +63% |
Adds .github/workflows/skill-review.yml that runs tesslio/skill-review
on PRs touching SKILL.md files. Posts one PR comment with scores and
feedback using only GITHUB_TOKEN — no Tessl login needed for contributors.
Non-blocking by default (fail-threshold: 0).
@rohan-tessl rohan-tessl marked this pull request as ready for review March 26, 2026 08:55
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant