skill-forge
Skill10 starsChanged 4 months agoMITView on GitHub
Add or evolve a /ratchet:* command across every synced surface at once — canonical prompt in PROMPTS.md, SKILL.md in the house shape, README row, templates, drift guard, tests. Use when a new loop move is being added to the plugin (like map or the probe were), or an existing skill's contract is changing. Prevents the four-surface drift the plugin-shape suite exists to catch.
49 other projects use a version of this.
Adopted by
Most recently taken up by
- brebuilds/claude-skillstheir own version, first seen Nov 8, 2026
- gxh98/haoge-skill-forgetheir own version, first seen Nov 8, 2026
- yashaswi-khurana/Skillstheir own version, first seen Nov 8, 2026
- Arcanea-Labs/Starlight-Intelligence-Systemtheir own version, first seen Nov 8, 2026
- AthenDrakomin-hub/agent-skillstheir own version, first seen Nov 1, 2026
- ligj1706/ProSkillstheir own version, first seen Nov 1, 2026
Used with
Projects with it also use sigma, agent-forge and 3 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
sanyuan0704Skillsigma
Personalized 1-on-1 AI tutor using Bloom's 2-Sigma mastery learning. Guides users through any topic with Socratic questioning, adaptive pacing, and rich visual output (HTML dashboards, Excalidraw concept maps, generated images). Use when user wants to learn something, study a topic, understand a concept, requests tutoring, says 'teach me', 'I want to learn', 'explain X to me step by step', 'help me understand', or invokes /sigma. Triggers on: learn, study, teach, tutor, understand, master, explain step by step.
Also in 3 projects that use this
xyreinsurance119Skillagent-forge
Создание, улучшение и аудит скиллов и агентов OpenClaw. Три режима: создание скилла (11 шагов), создание агента (9 шагов, с памятью и автоулучшением), улучшение существующего (5 шагов). Triggers: 'создай скилл', 'новый скилл', 'создай агента', 'новый агент', 'улучши скилл', 'skill creator', 'agent creator', 'скиллмейкер'.
Also in 3 projects that use this
localSummerSkillskill-review
回顾当前对话,整理 skill 使用情况、代码质量反馈和迭代建议,生成待确认的更新清单
Also in 3 projects that use this
NVIDIASkilldeep-research
Queue deep, multi-step research and analysis tasks to the DeepAgents worker. Supports execution depth presets, request-specific rubrics, SubAgent delegation, and RubricMiddleware cross-validation.
Also in 3 projects that use this
GoogleSkillcode-review
Adversarial systems, concurrency, memory-model, and architecture code review protocol. Use this skill whenever asked to review code, PRs, diffs, or branches in Filament.
Also in 3 projects that use this
Other versions49 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 49
- sanyuan0704/code-review-expert★ 3.9ksame name
- sanyuan0704/sanyuan-skills★ 3.9ksame name
- AgriciDaniel/skill-forge★ 166same name
- svy04/ballast★ 73same name
- danilokhury/Synabun★ 70same name
- motiful/skill-forge★ 45same name
- WilliamSaysX/skill-forge★ 43same name
- bjcoombs/ai-native-toolkit★ 32same name
- GodModeAI2025/skill-forge★ 18same name
- maroffo/claude-forge★ 17same name
39 more not listed
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.
In this repository24 other files it gives its agents.
The file
We split this skill into the rules it gives an agent
7 rules. 0 are common, 3 are distinctive, and 4 are very specific to this project.
0
Common
Most of the 42 edited versions we read keep these.
3
Distinctive
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 23
Read the two nearest-neighbor skills (for a gate-like command: skills/map/SKILL.md, skills/build/SKILL.md) — the new skill must read like …
- Very specific: line 28
Add the prompt as a fenced text block in reference/PROMPTS.md, in the imperative copy-paste voice of the existing twelve ("Produce the …
- Distinctive: line 31
Add a row to the Command ↔ prompt map table.
- Distinctive: line 32
If the mechanism is grafted from elsewhere: general mechanism, own words — rewrite in ratchet vocabulary and note provenance (aperture-dial …
- Very specific: line 85
Add /ratchet:<name> to the correct README section (Core loop / Specialized / Evolution) in the table style of its neighbors.
- Very specific: line 92
If the new command creates a cross-surface invariant the generic loops don't already cover (a template that must exist, a threading rule …
- Distinctive: line 104
All three suites green. Then:
“Common” is measured: kept by at least 6 in 10 of the 42 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
/skill-forge — one command, every surface
A /ratchet:* command is not a file; it is a contract spread across four synced
surfaces — reference/PROMPTS.md (intent), skills/<name>/SKILL.md (procedure),
README.md (catalog), and test/plugin-shape.test.js (drift police) — plus
templates/ when the command produces a file-shaped record. Editing one surface and
not the others is the repo's most common failure mode. This skill walks them in
dependency order: intent first, procedure second, catalog third, guard last.
Step 0 — Ground truth
ls skills/
grep -n "<command-name>" reference/PROMPTS.md README.md test/plugin-shape.test.jsRead the two nearest-neighbor skills (for a gate-like command: skills/map/SKILL.md,
skills/build/SKILL.md) — the new skill must read like it was written the same day.
Step 1 — Canonical prompt (PROMPTS.md) — the source of truth, written FIRST
- Add the prompt as a fenced
textblock inreference/PROMPTS.md, in the imperative copy-paste voice of the existing twelve ("Produce the artifact now." / "Map the fog before you build."). - Add a row to the Command ↔ prompt map table.
- If the mechanism is grafted from elsewhere: general mechanism, own words — rewrite in ratchet vocabulary and note provenance (aperture-dial precedent).
- Rule to preserve: the prompt must force pressure — a choice, an artifact, a test, a patch, serialized state, a killed option, or a push. If it only produces insight, it is not a ratchet command; stop and say so.
Step 2 — SKILL.md in the house shape
skills/<name>/SKILL.md, exactly this skeleton:
---
name: <name>
description: <verb-first pitch. Use when <trigger>. <what it produces / refuses>.>
argument-hint: "[...]" # only if it takes arguments
---
# /ratchet:<name> — <tagline>
<2–3 sentences: why this command exists, what failure it prevents.>
## Step 0 — Load stateratchet status
## Procedure
<numbered, imperative, bolded key rules. Prompt-level tripwires welcome, but never
claim they are enforced — enforcement is the CLI's job.>
## Output contract<FIXED-SHAPE block — same sections every run, emptiness stated>
## Serializeratchet <exact commands — artifact add / state append / state set phase>
<closing line naming the NEXT move — no ratchet skill ends in analysis.>Frontmatter description: is mandatory (plugin-shape asserts it) and doubles as the
when-to-use pitch the model router reads — write it like the existing ones.
Step 3 — Templates (only if the command produces a durable file)
Mirror templates/probe-card.md / templates/unknowns-map.md: terse field list,
every field load-bearing, no prose padding. Reference the template from the SKILL.md.
Step 4 — README row
Add /ratchet:<name> to the correct README section (Core loop / Specialized /
Evolution) in the table style of its neighbors. plugin-shape fails the suite if any
skill folder lacks a README mention — and if you RENAMED a command, grep README for
the old name; the suite also rejects stale names.
Step 5 — Drift guard
If the new command creates a cross-surface invariant the generic loops don't already
cover (a template that must exist, a threading rule like "build records deviations"),
add one ok(...) block to test/plugin-shape.test.js in the existing style: a
comment saying WHY the guard exists, then content-level asserts (/deviation/i.test(...)),
not snapshot dumps. Precedents: the map-wiring and probe-threading guards.
Step 6 — Verify + serialize
npm testAll three suites green. Then:
ratchet artifact add '{"title":"/ratchet:<name> command","kind":"code-patch","status":"v0","holes":[...]}'Output contract for this skill's own run:
SURFACES TOUCHED: PROMPTS.md § + map row · skills/<name>/SKILL.md · README § · [templates/…] · [plugin-shape guard]
SURFACES UNCHANGED AND WHY: <e.g. no template — command produces no durable file>
VERIFY: npm test <counts>
NEXT MOVE: /ratchet:attack the new SKILL.md before it ships<!-- Traced by: claude-fable-5 · 2026-07-07 -->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.

