agentleFS
steph-frtechsteph-frtech/claude-code-up/.claude/marketplaces/wshobson/CLAUDE.md

claude-code-up / wshobson

CLAUDE.md1 starsChanged 6 months agoMITView on GitHub

Claude Code plugin marketplace — 81 plugins (80 local + 1 external via git-subdir), 185 agents, 153 skills, 100 commands. Skill frontmatter: Only name is required. Agents, commands, and skills are auto-discovered from directory structure. Lists all plugin component paths for the registry. Agents as ./agents/name.md, skills as ./skills/skill-name (directory, not SKILL.md). Three-layer evaluation system in plugins/plugin-eval/. Full docs: docs/plugin-eval.md. triggering_accuracy (25%), orchestration_fitness (20%), output_quality (15%), scope_calibration (12%), progressive_disclosure (10%), token_efficiency (6%), robustness (5%), structural_completeness (3%), code_template_quality (2%), ecosystem_coherence (2%)

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

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

5 rules. About 0 are common, 1 is distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    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
  1. Very specific:

    Only name is required. Agents, commands, and skills are auto-discovered from directory structure.

    line 71
  2. Very specific:

    Use the Astral Rust toolchain: uv (package manager), ruff (linter/formatter), ty (type checker).

    line 146
  3. Very specific:

    Create plugins/<name>/ with .claude-plugin/plugin.json

    line 150
  4. Very specific:

    Add agents in agents/, commands in commands/, skills in skills/

    line 151
  5. Distinctive:

    Follow naming conventions: lowercase, hyphen-separated

    line 153

≈ “Common” is an estimate: at least 1 in 5 of the 27,641 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

Project: claude-agents

Claude Code plugin marketplace — 81 plugins (80 local + 1 external via git-subdir), 185 agents, 153 skills, 100 commands.

Repository Structure

claude-agents/
├── .claude-plugin/marketplace.json   # Registry of all plugins
├── plugins/                          # All 80 local plugins (1 more installed via git-subdir from external repo)
│   ├── <plugin-name>/
│   │   ├── .claude-plugin/plugin.json
│   │   ├── agents/*.md
│   │   ├── commands/*.md
│   │   └── skills/<skill-name>/SKILL.md
│   └── ...
├── docs/                             # Documentation
│   ├── plugins.md                    # Plugin catalog
│   ├── agents.md                     # Agent reference
│   ├── agent-skills.md               # Skills reference
│   ├── usage.md                      # Usage guide
│   ├── architecture.md               # Design principles
│   └── plugin-eval.md                # Evaluation framework
└── tools/                            # Development utilities

Plugin Authoring Conventions

Agent frontmatter

---
name: agent-name
description: "What this agent does. Use PROACTIVELY when [trigger conditions]."
model: opus|sonnet|haiku|inherit
color: blue|green|red|yellow|cyan|magenta # optional
tools: Read, Grep, Glob # optional — restricts available tools
---

Skill structure

skills/<skill-name>/
├── SKILL.md              # Required — frontmatter + content
├── references/           # Optional — supporting material
│   └── *.md
└── assets/               # Optional — templates, configs

Skill frontmatter:

---
name: skill-name
description: "Use this skill when [specific trigger conditions]."
---

Command frontmatter

---
description: What this command does
argument-hint: <path> [--flag]
---

plugin.json

Only name is required. Agents, commands, and skills are auto-discovered from directory structure.

{ "name": "plugin-name" }

marketplace.json

Lists all plugin component paths for the registry. Agents as ./agents/name.md, skills as ./skills/skill-name (directory, not SKILL.md).

Model Tiers

TierModelUse Case
Tier 1OpusArchitecture, security, code review, production coding
Tier 2InheritComplex tasks — user chooses model
Tier 3SonnetDocs, testing, debugging, support
Tier 4HaikuFast ops, SEO, deployment, simple tasks

PluginEval — Quality Evaluation Framework

Three-layer evaluation system in plugins/plugin-eval/. Full docs: docs/plugin-eval.md.

Quick Reference

cd plugins/plugin-eval

# Run tests
uv run pytest

# Evaluate a skill (static only)
uv run plugin-eval score path/to/skill --depth quick --output json

# Evaluate with LLM judge
uv run plugin-eval score path/to/skill --depth standard

# Full certification (all 3 layers)
uv run plugin-eval certify path/to/skill

# Compare two skills
uv run plugin-eval compare path/to/skill-a path/to/skill-b

# Build corpus for Elo ranking
uv run plugin-eval init plugins/

Evaluation Layers

  1. Static (Layer 1) — Deterministic structural analysis. < 2 seconds, free, always runs.
  2. LLM Judge (Layer 2) — Semantic evaluation via Claude (Haiku + Sonnet). ~30s, 4 LLM calls.
  3. Monte Carlo (Layer 3) — Statistical reliability via N simulated runs. ~2–5 min, 50–100 calls.

10 Dimensions (weights)

triggering_accuracy (25%), orchestration_fitness (20%), output_quality (15%), scope_calibration (12%), progressive_disclosure (10%), token_efficiency (6%), robustness (5%), structural_completeness (3%), code_template_quality (2%), ecosystem_coherence (2%)

Badges

Platinum ≥90, Gold ≥80, Silver ≥70, Bronze ≥60

Anti-Patterns

OVER_CONSTRAINED (>15 MUST/ALWAYS/NEVER), EMPTY_DESCRIPTION (<20 chars), MISSING_TRIGGER (no "Use when…"), BLOATED_SKILL (>800 lines no refs), ORPHAN_REFERENCE (dead link), DEAD_CROSS_REF (missing skill)

Tech Stack

  • Python ≥ 3.12, uv, ruff, ty, pytest
  • Dependencies: pydantic, typer, rich, pyyaml
  • Optional: claude-agent-sdk (LLM layers), anthropic (API alternative)

Development

Python Tooling

Use the Astral Rust toolchain: uv (package manager), ruff (linter/formatter), ty (type checker). Do not use pip, mypy, or black.

Adding a Plugin

  1. Create plugins/<name>/ with .claude-plugin/plugin.json
  2. Add agents in agents/, commands in commands/, skills in skills/
  3. Update .claude-plugin/marketplace.json
  4. Follow naming conventions: lowercase, hyphen-separated

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.