AgriciDanielAgriciDaniel/skill-forge/CLAUDE.md

skill-forge

CLAUDE.md166 starsChanged 8 months agoMITView on GitHub

This repository contains Skill Forge, a Tier 4 Claude Code skill for creating, reviewing, evolving, and publishing other Claude Code skills. It follows the Agent Skills open standard and the 3-layer architecture (directive, orchestration, execution).

1 other project uses a version of this.

Adopted by

Most recently taken up by

lovepixel-git/skill-forgeword for word, first seen Aug 1, 2026

Other versions1 project uses a version of this.
Word for word · 1
lovepixel-git/skill-forge★ 0100% same
Edited · 0
None found.
Wrote their own version · 0
None found.
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 repositoryOne other file it gives its agents.
- skill-forgeSkill

The file

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

5 rules. About 0 are common, 3 are distinctive, and 2 are very specific to this project.

≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.

Common
Distinctive
Very specific

Distinctive:
Simplicity First: Start with Tier 1, evolve to higher tiers only when needed
line 69
Very specific:
Test with python skill-forge/scripts/validate_skill.py <path> after changes
line 73
Very specific:
Keep SKILL.md files under 500 lines / 5000 tokens
line 74
Distinctive:
Follow kebab-case naming for all skill directories
line 77
Distinctive:
Never put README.md inside skill folders
line 78

≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.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

CLAUDE.md

Skill Forge — Ultimate Claude Code Skill Creator

Project Overview

This repository contains Skill Forge, a Tier 4 Claude Code skill for creating, reviewing, evolving, and publishing other Claude Code skills. It follows the Agent Skills open standard and the 3-layer architecture (directive, orchestration, execution).

Architecture

skill-forge/                         # Repository root
  CLAUDE.md                          # Project instructions
  skill-forge/                       # Main orchestrator skill (Tier 4)
    SKILL.md                         # Entry point, routing table, core rules
    references/                      # On-demand knowledge files (10 files)
      anatomy.md                     # Skill file structure, naming, agent format
      patterns.md                    # Proven workflow patterns
      frontmatter-spec.md           # YAML frontmatter specification
      description-guide.md          # Writing effective descriptions
      testing-guide.md              # Testing methodology
      pro-agent.md                  # 3-layer architecture deep dive
      tools-reference.md            # Tool names, permission patterns, MCP
      hooks-reference.md            # Hook events, types, quality gates
      skills-activation.md          # Skill discovery, activation, features
      platforms.md                   # Multi-platform conversion rules
    scripts/                         # Deterministic execution scripts
      init_skill.py                 # Scaffold new skills
      validate_skill.py             # Validate skill structure
      package_skill.py              # Package for distribution
      convert_skill.py              # Convert skills to other platforms
      generate_eval_set.py          # Generate trigger eval sets from SKILL.md
      aggregate_benchmark.py        # Aggregate eval results into benchmarks
      optimize_description.py       # Optimize descriptions with train/test split
      skill_utils.py                # Shared utilities (frontmatter parser)
    assets/
      templates/                     # Skill templates by tier
        minimal.md                  # Tier 1: single SKILL.md
        workflow.md                 # Tier 2: SKILL.md + scripts
        multi-skill.md             # Tier 3: orchestrator + sub-skills
        ecosystem.md               # Tier 4: full ecosystem
  skills/                            # Sub-skills
    skill-forge-plan/SKILL.md       # Architecture and design planning
    skill-forge-build/SKILL.md      # Scaffold and generate skills
    skill-forge-review/SKILL.md     # Audit and validate skills
    skill-forge-evolve/SKILL.md     # Improve skills from feedback
    skill-forge-eval/SKILL.md       # Run eval pipeline with assertions
    skill-forge-benchmark/SKILL.md  # Benchmark performance tracking
    skill-forge-publish/SKILL.md    # Package and distribute skills
    skill-forge-convert/SKILL.md    # Convert skills to other platforms
  agents/                            # Subagent definitions
    skill-forge-architect.md        # Architecture design agent
    skill-forge-writer.md           # SKILL.md content writer agent
    skill-forge-validator.md        # Validation agent
    skill-forge-converter.md        # Platform conversion agent
    skill-forge-executor.md         # Eval execution agent
    skill-forge-grader.md           # Eval grading agent
    skill-forge-analyzer.md         # Benchmark analysis agent
    skill-forge-comparator.md       # Blind A/B comparison agent
  install.sh                         # Installation script

Key Principles

Progressive Disclosure: Metadata always loaded, instructions on activation, resources on demand
Description is King: The YAML description field determines when skills activate
3-Layer Architecture: Directives (SKILL.md) + Orchestration (Claude) + Execution (scripts)
Self-Annealing: Learn from failures, update skills with discoveries
Simplicity First: Start with Tier 1, evolve to higher tiers only when needed

Development Rules

Test with python skill-forge/scripts/validate_skill.py <path> after changes
Keep SKILL.md files under 500 lines / 5000 tokens
Reference files should be focused and under 200 lines
Scripts must have docstrings, CLI interface, JSON output
Follow kebab-case naming for all skill directories
Never put README.md inside skill folders

Commands

Command	Purpose
`/skill-forge`	Interactive skill creation wizard
`/skill-forge plan`	Design skill architecture
`/skill-forge build`	Scaffold and generate skill files
`/skill-forge review`	Audit existing skill quality
`/skill-forge evolve`	Improve skill from feedback
`/skill-forge eval`	Run eval pipeline to test skill quality
`/skill-forge benchmark`	Benchmark skill performance
`/skill-forge publish`	Package for distribution
`/skill-forge convert`	Convert skills to other platforms

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