agentleFS
SkillMedevSkillMedev/skill-builder-studio/skills/skill-creator/SKILL.md

skill-creator

Skill0 starsChanged 39 days agoMITView on GitHub

Use when authoring a brand-new skill from scratch - turning a capability idea or a \"make me a skill that…\" request into a complete SKILL.md with a trigger-precise description and the full paid-quality anatomy: procedure, elicitation, thresholds, worked artifact, deliverable, Do NOT, quality bar. Triggers on \"write a new SKILL.md from scratch\", \"turn this capability idea into a complete skill\". Do NOT use to review or grade an existing skill (use skill-auditor), to convert an existing prompt (use prompt-to-skill), or for writing unrelated to skills.

189 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use taskflow-inbox-triage, mermaid-diagram-generator and 10 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.
  • AtomicBot-aiSkill

    taskflow-inbox-triage

    Also in 6 projects that use this

  • shalombSkill

    mermaid-diagram-generator

    Guide on creating and editing Mermaid charts safely. Use when asked to draw a diagram, flowchart, or architecture chart. Triggers on: 'mermaid', 'diagram', 'flowchart', 'visualize'.

    Also in 5 projects that use this

  • naman10parikhSkill

    ecosystem-reference

    Quick-access directory of Claude Code ecosystem repos, skills, plugins, MCPs, and patterns. Use when improving the harness.

    Also in 5 projects that use this

  • naman10parikhSkill

    fractal-delegation

    Also in 5 projects that use this

  • Liang-MHSkill

    stock_analysis

    获取 A 股、基金和港股的实时行情和技术分析数据

    Also in 5 projects that use this

  • nicepkgSkill

    community-led-growth

    Expert in community-led growth (CLG) - leveraging user communities to drive acquisition, retention, and expansion. Covers building developer communities, user groups, ambassador programs, and turning customers into advocates. Knows the difference between community as a feature vs community as a growth engine, and how to measure community ROI. Use when "community-led, CLG, ambassador program, champion program, user community, developer community, word of mouth, user group, " mentioned.

    Also in 5 projects that use this

  • +6 more from naman10parikh

The file

We split this skill into the rules it gives an agent

7 rules. 0 are common, 7 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 185 edited versions we read keep these.

  • 7

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Check for collisions: Search the catalog (browse_skills) for skills whose triggers could fire on the same request.

    line 26
  2. Distinctive:

    Do not write the body before the trigger is nailed.

    line 44
  3. Distinctive:

    Do not ship an essay. If the body has no procedure, no numbers, and no artifact, it does not clear the bar - get the substance or say what …

    line 45
  4. Distinctive:

    Do not pad: no preamble, no hedging, no restating what the model already knows.

    line 46
  5. Distinctive:

    Do not ship territory collisions.

    line 47
  6. Distinctive:

    Do not put dates, versions, or "latest" anywhere in the file.

    line 48
  7. Distinctive:

    Do not grade existing skills here - that is skill-auditor's job.

    line 49

“Common” is measured: kept by at least 6 in 10 of the 185 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Skill Creator

Turn a capability idea into a committable SKILL.md that meets the paid-quality bar. The costly mistake this prevents: writing a competent essay about the topic instead of an operating procedure - a skill that reads well and changes nothing about the model's output.

The bar

The skill you write ships only if all of these hold:

  • It fires on exactly the right requests and never on the wrong ones. The description front-loads concrete triggers a router can act on with no ambiguity.
  • It does not collide with any existing skill. Before writing, check the catalog for trigger overlap and carve clean boundaries into both descriptions.
  • Its body carries the full anatomy (skill-author is the canonical standard): framing, ordered procedure, input elicitation, concrete thresholds, at least one worked artifact, an explicit deliverable, a Do NOT section, and a quality bar.
  • It is token-disciplined. The description loads every session; the body loads on trigger. Every line is paid for by every installer, every session.
  • It is correct and self-contained - real tools, real paths, real practitioner numbers; no hallucinated capabilities; no context it hasn't established.

Operating procedure

  1. Find the trigger first. Before any body text, answer: what exact request should make this fire? Name concrete signals - task verbs, named artifacts, file types, the phrases a user would actually type. If you cannot name them precisely, the skill is not ready to write; press for the real use case.
  2. Check for collisions. Search the catalog (browse_skills) for skills whose triggers could fire on the same request. Carve the boundary and write it into both descriptions as negative scope. A new skill that steals a neighbor's territory is a defect.
  3. Gather the domain substance. Collect the frameworks, thresholds, and templates the skill will encode - from the requester, or from established practitioner methods you name. A skill with no numbers and no artifact is an essay; if the substance is not available yet, say what is missing rather than padding around it.
  4. Write the description - trigger first. WHAT + WHEN + NOT (skill-description-writer is the canonical standard): open with when to activate, quoting real user phrasing; state what it produces; append "Do NOT use for X - use <slug> instead" wherever a misfire into a named neighbor is plausible.
  5. Write the body in anatomy order. Framing paragraph (outcome + the mistake prevented), numbered procedure with dependency reasoning, input elicitation with defaults, thresholds with real numbers, worked artifact (filled example, good/bad pair, FILL-field template, or runnable calculator with expected output shown - fenced code blocks are correct here), Deliverable, Do NOT with the why, quality bar, and an escalation boundary where the domain touches legal, medical, or financial ground.
  6. Set name and slug. Name in Title Case; slug in kebab-case matching the dominant noun. Match the house voice of the existing catalog so the set reads as one author.
  7. Self-check, then hand to skill-auditor. Re-read the description cold: fires on the right request, silent on a near-miss? Re-read the body: could a practitioner execute without a clarifying question? Is there an artifact they could copy out and use as-is? Then state the skill is ready for a skill-auditor pass before publishing.

Deliverable

A single committable SKILL.md - valid frontmatter (name, description) plus the full-anatomy body - at the correct slug/SKILL.md path. No commentary baked into the file.

Do NOT

  • Do not write the body before the trigger is nailed. A vague trigger produces a skill that never fires or always fires, however good the body.
  • Do not ship an essay. If the body has no procedure, no numbers, and no artifact, it does not clear the bar - get the substance or say what is missing.
  • Do not pad: no preamble, no hedging, no restating what the model already knows. If a line does not change behavior, cut it.
  • Do not ship territory collisions; fix the boundary in both descriptions before delivering.
  • Do not put dates, versions, or "latest" anywhere in the file.
  • Do not grade existing skills here - that is skill-auditor's job; hand off rather than absorbing its scope.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.