agentleFS
Sign inStart free
h3y6eh3y6e/spec-skills/AGENTS.md

spec-skills

AGENTS.md0 starsChanged 9 months agoMITView on GitHub

spec-skills is a Markdown-only repository containing Agent Skills for Spec-Driven Development. There is no source code, no build system, no package manager, and no test framework. All content is declarative Markdown defining AI agent skill definitions and their reference templates. There are no build, lint, or test commands. The only CI pipeline is automated release, triggered on push to main. It bumps the metadata.version field in all five SKILL.md files and creates a GitHub release. Every skill file follows this…

  • Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.

    Of the AGENTS.md files we read, the 31,774 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.

  • In this repository6 other files it gives its agents.

The file

We split this AGENTS.md into the rules it gives an agent

10 rules. About 0 are common, 9 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈9

    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
  1. Distinctive:

    All content is declarative Markdown defining AI agent skill definitions and their reference templates.

    line 7
  2. Distinctive:

    Every skill file follows this exact structure:

    line 31
  3. Distinctive:

    Use numbered lists for sequential steps

    line 39
  4. Very specific:

    Use - [ ] / - [x] checkboxes for DoD items and acceptance criteria

    line 40
  5. Distinctive:

    Use backticks for file paths, identifiers, and skill names

    line 42
  6. Distinctive:

    Keep all documents within 150 lines when possible

    line 44
  7. Distinctive:

    Keep descriptions to 1-3 sentences

    line 63
  8. Distinctive:

    Avoid redundancy; deduplicate during final review

    line 64
  9. Distinctive:

    Every skill ends with Completion Guidance recommending the next skill

    line 65
  10. Distinctive:

    Follow conventional commits with lowercase type prefixes:

    line 71

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md 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.

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

AGENTS.md

Project Overview

spec-skills is a Markdown-only repository containing Agent Skills for Spec-Driven Development. There is no source code, no build system, no package manager, and no test framework. All content is declarative Markdown defining AI agent skill definitions and their reference templates.

Commands

There are no build, lint, or test commands.

The only CI pipeline is automated release, triggered on push to main. It bumps the metadata.version field in all five SKILL.md files and creates a GitHub release.

Workflow Skills

  1. spec-constitution - Define shared project rules in specs/constitution.md when the team wants durable workflow governance
  2. spec-specify - Create specs/{feature}/spec.md from requirements
  3. spec-plan - Create specs/{feature}/plan.md as the execution document with key decisions, phased tasks, and DoD
  4. spec-implement - Execute tasks with implement/self-validate/fix loops

Utility Skills

  • spec-research - Investigate a topic and record findings in specs/{feature}/research/{topic}.md; callable from any phase or standalone

Content Style Guidelines

SKILL.md Structure

Every skill file follows this exact structure:

  1. YAML frontmatter: name, description, license, metadata.author, metadata.version
  2. H1 title: # {Skill Name} Skill
  3. Sections in order: Purpose, Input, Output, Steps, Success Criteria, Completion Guidance

Markdown Formatting

  • Use numbered lists for sequential steps
  • Use - [ ] / - [x] checkboxes for DoD items and acceptance criteria
  • Task IDs follow the pattern T001, T002, etc.
  • Use backticks for file paths, identifiers, and skill names
  • Requirements use prefixed IDs: FR-001, SC-001
  • Keep all documents within 150 lines when possible

Templates

  • Templates live in skills/{skill-name}/references/
  • Placeholders use [UPPER_SNAKE_CASE] format (e.g., [FEATURE_NAME], [DESCRIPTION])
  • Templates include placement instructions (e.g., Place this file at specs/{feature}/spec.md.)
  • HTML comments (<!-- ... -->) are used for inline guidance in templates

Naming Conventions

  • Skill directories: spec-{name} (kebab-case with spec- prefix)
  • Feature directories: specs/{feature-slug}/ (kebab-case)
  • Research files: specs/{feature}/research/{topic}.md (kebab-case)
  • Fixed filenames: SKILL.md, spec.md, plan.md; when used, the constitution file is constitution.md

Writing Style

  • Concise, imperative tone in skill steps ("Read", "Create", "Update", "Perform")
  • Keep descriptions to 1-3 sentences
  • Avoid redundancy; deduplicate during final review
  • Every skill ends with Completion Guidance recommending the next skill

Git Conventions

Commit Messages

Follow conventional commits with lowercase type prefixes:

  • feat: for new features or skills
  • fix: for corrections
  • docs: for documentation changes
  • chore: for maintenance (config, CI)
  • ci: for CI/workflow changes

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