agentleFS
Sign inStart free
DROOdotFOODROOdotFOO/agent-skills/skills/skill-creator/SKILL.md

skill-creator

Skill1 starsChanged 2 months agoMITView on GitHub

Interactively scaffold new agent skills with correct frontmatter, trigger clauses, sub-files, and linter compliance. TRIGGER when: user asks to create a new skill, scaffold a skill, add a skill to agent-skills, or says "new skill" or "skill creator". DO NOT TRIGGER when: user is editing an existing skill, writing AGENTS.md or CLAUDE.md instructions, or building an MCP server (use mcp-server-builder skill).

We split this skill into the rules it gives an agent

2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 183 edited versions we read keep these.

  • 2

    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 overlap with existing skills before creating

    line 39
  2. Distinctive:

    After the frontmatter, structure the body:

    line 65

“Common” is measured: kept by at least 6 in 10 of the 183 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

Scaffold new agent skills that pass ./scripts/skills-lint.sh on the first try. Walk the user through naming, trigger design, sub-file structure, and quality review.

What is a skill?

A skill is a directory under skills/<name>/ containing markdown files with YAML frontmatter. The entry point is always SKILL.md. Skills provide progressively disclosed guidance for compatible hosts such as Codex, Claude Code, and Raxol. Hosts use the description to decide when the skill applies and load the full instructions only when needed.

Skills are NOT agents, plugins, or code libraries. They are structured knowledge documents that shape an agent host's behavior for a specific domain.

Scaffolding workflow

Step 1: Choose a name

  • Lowercase, hyphenated: my-skill-name
  • Short but specific: prefer database-designer over db or database-schema-design-and-optimization-tool
  • Check for overlap with existing skills before creating

Step 2: Write trigger clauses

The description field in SKILL.md frontmatter MUST include both:

  • TRIGGER when: -- enumerate concrete conditions (imports, file types, user phrases)
  • DO NOT TRIGGER when: -- exclude adjacent domains to prevent false activation

Good triggers are specific and observable:

TRIGGER when: user asks to design a REST API, define endpoint schemas,
or working with OpenAPI spec files (.yaml/.json with openapi: field).
DO NOT TRIGGER when: user is implementing API handlers (use the
language-specific skill), or configuring an API gateway.

Bad triggers are vague:

TRIGGER when: user is working with APIs.

Step 3: Write SKILL.md body

After the frontmatter, structure the body:

  1. One-paragraph summary -- what the skill does, when to use it
  2. Workflow or phases -- the step-by-step process the skill guides
  3. Key rules or checklists -- what to always/never do
  4. Common pitfalls -- table of mistake/fix pairs
  5. Reading guide -- table linking to sub-files with backtick-wrapped paths

Step 4: Create sub-files

Each sub-file covers one focused topic. Every sub-file needs YAML frontmatter with three required fields:

  • impact: one of CRITICAL, HIGH, MEDIUM, LOW
  • impactDescription: one sentence explaining why this matters
  • tags: comma-separated string (NOT a YAML array)

Sub-files should contain incorrect/correct example pairs showing what to do and what not to do, with brief explanations.

Step 5: Validate

./scripts/skills-lint.sh

The linter checks:

  • SKILL.md exists with name:, description:, metadata: fields
  • Description contains TRIGGER when: and DO NOT TRIGGER
  • Sub-files have YAML frontmatter with impact and impactDescription
  • Tags are comma-separated strings, not YAML arrays
  • File references in reading guide tables point to real files
  • Cross-skill references in "See also" point to real skill directories

What You Get

  • A complete skills/<name>/ directory with a SKILL.md entry point containing valid frontmatter and trigger clauses
  • Sub-files with correct YAML frontmatter (impact, impactDescription, tags) and incorrect/correct example pairs
  • A skill that passes ./scripts/skills-lint.sh on the first run with no manual fixes needed

Reading guide

TopicFile
Copy-paste frontmatter templates for SKILL.md and sub-filestemplates.md
Quality checklist, anti-patterns, review criteriaquality-checklist.md
Verification-before-completion pattern for workflow skillsverification.md

See also

  • codebase-onboarding
  • architect

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.

Check your setup, free