skill-map
Skill4 starsChanged 3 months agoMITView on GitHub
Turn the whole skill library on this machine into living HTML guide pages plus a skills-map.csv, covering every skill's wake words, which agent homes have it (Claude Code, Cursor, Codex, Gemini), usage counts, trigger conflicts, and created vs downloaded. Use when the user says skill map, map my skills, audit my skills, skill usage, skill conflicts, skill drift, or after installing any batch of new skills. Cards read live from each SKILL.md frontmatter so the map never drifts from reality.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- tdfydfy/skill-maptheir own version, first seen Jul 19, 2026
- LJMStark/geo-doctortheir own version, first seen Jul 3, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- LJMStark/geo-doctor★ 0same name
- tdfydfy/skill-map★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 19
Run it from this skill's folder.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Skill Map
One command regenerates every guide page from the real skill files. The description in each SKILL.md IS the auto-trigger contract, so rendering descriptions onto cards makes the map self-truthing: if a card is wrong, the skill's own trigger is wrong, and that is a finding, not a map bug.
Regenerate
python generate.pyRun it from this skill's folder. Output lands in ./skill-map-out/ (index.html plus one page per suite) unless manifest.json sets output_dir. Open index.html for the user when done.
The script prints a SKIPPED list for any manifest name missing on disk. A skip means the manifest or the skill folder changed: fix the name in manifest.json or accept the retirement by removing the entry. Completion: rerun until SKIPPED is empty or every remaining skip is intentional and reported.
Zero-config first run
Out of the box, manifest.json ships empty and the map runs in auto mode: every skill found in ~/.claude/skills renders into one suite called "Your skills". Nothing to configure, nothing to fill in.
Three optional keys in manifest.json unlock curation:
suites: group skills into named suites, each{slug, title, tagline, accent, skills}. The moment this is non-empty, auto mode turns off and only listed skills get suite pages.output_dir: absolute path for the generated site, when you want it somewhere permanent.debriefs_dir: a folder of session notes (markdown) to mine for usage mentions. Unset, usage counts come from agent transcripts alone.
Add or move skills
Suites, accents, and membership live in manifest.json. One skill can appear in two suites when it genuinely serves both. After any skill batch ships, add the new names to the right suites and regenerate; the batch is not done until the map shows it.
Design contract
Pages follow da-edu-html conventions: #0A0B0D background, #12131A cards, per-suite accent from the DA palette, Space Grotesk + JetBrains Mono, staggered card entrances, mobile responsive, digitalalchemy.dev footer CTA. Card anatomy: /slash-name, an auto-triggers or slash-only badge (read from disable-model-invocation), what-it-does, and FIRES WHEN wake words split out of the real description. Change the design in generate.py once; every page inherits it.
Scope rule
The suite pages cover the user's own skills. That boundary is the manifest's job: skills you installed from elsewhere never get manifest entries; they appear on downloaded.html and in the CSV with origin downloaded, automatically, by set difference against the manifest. (In auto mode everything in ~/.claude/skills counts as yours; curate suites when you want the line drawn precisely.)
v2 data sources (what each signal means)
- Usage = mentions across your
debriefs_dirnotes (when set in manifest.json) + real invocations grepped from~/.claude/projectstranscripts ("Launching skill:" and command-name markers, via ripgrep when available, with a quiet skip when not). Debrief mentions overcount slightly (a skill discussed is counted); invocations undercount other agents. Read them as a ranking, not an audit. - Conflicts = two detectors: name clashes against locally cached plugin skills, and trigger-word overlap pairs (Jaccard on FIRES WHEN tokens, threshold 0.30 or 8 shared words). An overlap pair means the two descriptions can steal each other's wake words; fix by sharpening the weaker description, not by deleting a skill.
- skills-map.csv = every mapped skill plus every downloaded skill, one row each, Excel-safe utf-8-sig. Columns include suite, origin, invocation type, homes, usage split, and conflict flags.
Changelog
- 2.1.0: builder profile + effectiveness grades (fair NEW handling via file mtime, adoption rescaled), sticky suite nav on every page, grade/usage filter buttons + sort control, dedicated conflicts.html with shared wake words per pair, run-over-run history deltas (history.jsonl), common-word usage-count fix, CSV adds grade/score/breakdown/last_touched/new columns.
- 2.0.0: usage counts, conflict detection, created-vs-downloaded page, per-home presence badges, CSV export.
- 1.0.0: suite pages + index from live frontmatter.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

