SkillForge
Indexes your installed skills and triages any input into use / improve / create
Test report
- Verdict
- Works with setup
- Score
- Tested
- Jul 16, 2026
- Environment
- Claude Code 2.x (agent harness)
- Upstream re-checked
- Jul 29, 2026 · e20fab4
⚠ This skill is no longer available upstream. Our re-check on Aug 10, 2026 couldn't find it any more (SKILL.md not found in repo tree). The test below is what we measured on Jul 16, 2026 and we're leaving it up as a record — but there is nothing left to install, so we've removed the command.
Task on a real directory of 469 installed skills: 'which of my skills translates blog posts into other languages for SEO?'. Baseline run (grep -rl 'translat' on ~/.claude/skills/*/SKILL.md) yielded 32 hits, the correct answer seo-translator was in 20th position without any ranking. The skill worked properly: discover_skills.py indexed 469 skills in 0.87 s, triage_skill_request.py raised seo-translator to 2nd place — but with 20% against generic 'seo' at 35% (match only by substring in name) and returned CLARIFY instead of USE_EXISTING, although the exact skill is in the system. Worse: for 'do I have a skill for database migrations', it confidently recommends b2b-local-outreach at 82% — simply because the word 'database' appears in the description; for 'help me debug this TypeError' — chatcrystal-debug-recall at 96% by substring 'debug'. Its own documented examples are not reproducible: SKILL.md promises routing to debug skills for 'TypeError: Cannot read property 'map'', but we actually get CREATE_NEW with a top match of 31%. Validators measure not quality, but conformity to their template: quick_validate.py issued 'Skill is valid!' for a fresh skeleton with 21 TODOs in the body and for a skill with a broken link to a non-existent references/nope.md, while validate-skill.py scored SkillForge itself 82/92, but working symbiotic-onboard 8/16 and fable-domain 6/15. Pros honestly exist: 16 attached pytest tests pass, scaffolder and index work, request type classifier is correct in 5 out of 6 cases.
Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works
- Installs cleanly 4/5
- Triggers reliably 4/5
- Output vs. baseline 5/10
- Docs & honesty 2/5
What SkillForge does
Builds a searchable index of every skill installed across ~/.claude, ~/.codex and ~/.agents, then classifies any input and routes it to an existing skill, an improvement, or a new skill scaffold. Also ships a proactive Context Skill Advisor that can run on a local schedule. Triggers on "SkillForge: <goal>", "do I have a skill for...", "which skill", and "improve <skill> skill".
How to install SkillForge
Nothing to install: the source repository no longer has this skill. If the author brings it back, our daily re-check will pick it up and the command will reappear here.
Commands — how to trigger SkillForge
-
/skillforgeIndexes your installed skills and triages any input into use / improve / create
It also activates on plain-language prompts like these:
-
Do I have a skill installed that handles SEO translation of blog posts -
Which of my skills should I use to fix this TypeError I'm debugging -
Scaffold a brand new skill since nothing I have covers database migrations
Frequently asked questions
- Is the SkillForge skill free?
- Yes. The skill itself is free from tripleyak/SkillForge. SkillProof publishes the install command and an independent test verdict at no cost.
- Does SkillForge work with Claude Code?
- We tested it with Claude Code 2.x (agent harness) on Jul 16, 2026. Verdict: Works with setup. Task on a real directory of 469 installed skills: 'which of my skills translates blog posts into other languages for SEO?'. Baseline run (grep -rl 'translat' on ~/.claude/skills/*/SKILL.md) yielded 32 hits, the correct answer seo-translator was in 20th position without any ranking. The skill worked properly: discover_skills.py indexed 469 skills in 0.87 s, triage_skill_request.py raised seo-translator to 2nd place — but with 20% against generic 'seo' at 35% (match only by substring in name) and returned CLARIFY instead of USE_EXISTING, although the exact skill is in the system. Worse: for 'do I have a skill for database migrations', it confidently recommends b2b-local-outreach at 82% — simply because the word 'database' appears in the description; for 'help me debug this TypeError' — chatcrystal-debug-recall at 96% by substring 'debug'. Its own documented examples are not reproducible: SKILL.md promises routing to debug skills for 'TypeError: Cannot read property 'map'', but we actually get CREATE_NEW with a top match of 31%. Validators measure not quality, but conformity to their template: quick_validate.py issued 'Skill is valid!' for a fresh skeleton with 21 TODOs in the body and for a skill with a broken link to a non-existent references/nope.md, while validate-skill.py scored SkillForge itself 82/92, but working symbiotic-onboard 8/16 and fable-domain 6/15. Pros honestly exist: 16 attached pytest tests pass, scaffolder and index work, request type classifier is correct in 5 out of 6 cases.
- What is the SkillForge SkillProof Score?
- 6.0/10 — installs cleanly 4/5, triggers reliably 4/5, output vs. baseline 5/10, docs & honesty 2/5.
- How do I install SkillForge?
- Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
- Can I use SkillForge with Cursor, Copilot, Gemini CLI, Codex or other AI tools?
- The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.