agentleFS
4444J994444J99/a-i--skills/CLAUDE.md

a-i--skills

CLAUDE.md17 starsChanged 5 months agoApache-2.0Archived repositoryView on GitHub

This is Anthony Padavano's (4444J99) agent skills repository—a catalog of skills that extend AI coding agents' capabilities. Each skill is a self-contained folder with a SKILL.md file containing YAML frontmatter and instructions. Skill locations: - Catalog skills: Located in skills/ directory, organized by category (e.g., skills/creative/algorithmic-art/, skills/development/mcp-builder/). 167 skills across 12 categories. - Plugin skills: Claude Code plugins under plugins/ (11 skills).

1 other project uses a version of this.

The file

Show its rules and text

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

13 rules. About 0 are common, 4 are distinctive, and 9 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 9

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    All skills must have a license field (MIT for open skills)

    line 142
  2. Distinctive:

    Only redistributable (open-licensed) skills belong here.

    line 144
  3. Very specific:

    Update docs/THIRD_PARTY_NOTICES.md when adding external assets

    line 146
  4. Distinctive:

    Use GitHub MCP tools ONLY for operations that require the GitHub API:

    line 153
  5. Very specific:

    If .conductor/active-handoff.md exists, READ IT FIRST before doing any work.

    line 179
  6. Distinctive:

    If the handoff says "CROSS-VERIFICATION REQUIRED", your self-assessment will NOT be trusted.

    line 183
  7. Distinctive:

    NOT be trusted. A different agent will verify your output against these constraints.

    line 184
  8. Very specific:

    Run organvm session review --latest to get a session summary

    line 189
  9. Very specific:

    Check for unimplemented plans: organvm session plans --project .

    line 190
  10. Very specific:

    Run organvm prompts distill --dry-run to detect uncovered operational patterns

    line 192
  11. Very specific:

    Run organvm atoms pipeline --write && organvm atoms fanout --write to generate task queue.

    line 243
  12. Very specific:

    Orchestration: Use conductor patch for system status and work queue.

    line 303
  13. Very specific:

    Lifecycle: Follow the FRAME -> SHAPE -> BUILD -> PROVE workflow.

    line 304

≈ “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

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Repository Overview

This is Anthony Padavano's (4444J99) agent skills repository—a catalog of skills that extend AI coding agents' capabilities. Each skill is a self-contained folder with a SKILL.md file containing YAML frontmatter and instructions.

Skill locations:

  • Catalog skills: Located in skills/ directory, organized by category (e.g., skills/creative/algorithmic-art/, skills/development/mcp-builder/). 167 skills across 12 categories.
  • Plugin skills: Claude Code plugins under plugins/ (11 skills).

Repository Structure

ai-skills/
├── README.md              # Repository documentation
├── CLAUDE.md              # Claude Code instructions (this file)
├── skills/                # All example skills (167), organized by category
│   ├── creative/          # Art, music, design (15 skills)
│   ├── data/              # Data analysis and ML (8 skills)
│   ├── development/       # Coding patterns and tools (47 skills, incl. bundles)
│   ├── documentation/     # Docs and GitHub profiles (7 skills)
│   ├── education/         # Teaching and learning (6 skills)
│   ├── integrations/      # Third-party integrations (14 skills)
│   ├── knowledge/         # Knowledge management (7 skills)
│   ├── professional/      # Business and career (12 skills)
│   ├── project-management/ # Planning and roadmaps (6 skills)
│   ├── security/          # Security and compliance (6 skills, incl. bundle)
│   ├── specialized/       # Niche domains (6 skills)
│   └── tools/             # Meta-skills and orchestration (11 skills)
├── plugins/               # Claude Code plugins (11 skills)
├── docs/                  # Documentation files
│   ├── CHANGELOG.md
│   ├── CONTRIBUTING.md
│   ├── ROADMAP.md
│   ├── api/               # Skill spec, federation schema, activation conditions
│   └── guides/            # Getting started, overrides, core-vs-community
├── agents/                # AI agent definitions
│   └── skill-planner.md   # Skill composition planner
├── commands/              # Slash commands
│   ├── plan-workflow.md   # /plan-workflow - generate skill chains
│   └── skill-health.md   # /skill-health - run health checks
├── scripts/               # Build and validation tools
├── staging/               # Skills in development
├── config/                # Authored runtime catalog configuration
├── distributions/         # Generated copies retained during installer migration
│   ├── collections/       # Skill path lists + tier lists
│   ├── skills-registry.json  # Machine-readable skill metadata
│   ├── skills-lock.json   # Lockfile with SHA-256 hashes
│   ├── claude/            # Claude Code bundles
│   ├── codex/             # Codex bundles
│   ├── direct/            # Direct link directories
│   └── extensions/        # Gemini CLI extensions
└── .claude-plugin/        # Marketplace metadata

Common Commands

# Build from canonical source and activate a complete verified release
python3 scripts/skills_install.py install
python3 scripts/skills_install.py path

# Refresh collections, registry, lockfile, and generated link directories
python3 scripts/refresh_skill_collections.py
python3 scripts/refresh_skill_collections.py --mode symlink  # Use symlinks instead of copies

# Validate skill frontmatter
python3 scripts/validate_skills.py --collection example --unique --check-links

# Run the tooling test suite
python3 -m pytest

# Verify generated bundles, registry, and lockfile are in sync
python3 scripts/validate_generated_dirs.py

# Run skill health checks (scripts, references, size metrics)
python3 scripts/skill_health_check.py                        # All skills
python3 scripts/skill_health_check.py --skill mcp-builder    # Single skill
python3 scripts/skill_health_check.py --json                 # JSON output

# PR validation report (for CI)
python3 scripts/pr_validation_report.py --base origin/main

# Full release workflow (bumps versions, updates changelog, commits, tags, pushes)
python3 scripts/release.py 1.2.0 \
  --change "Description" \
  --add "New feature" \
  --fix "Bug fix" \
  --commit --tag --push --release --notes-from-changelog

Architecture

Skill Structure

my-skill/
  SKILL.md          # Required: YAML frontmatter + markdown instructions
  scripts/          # Optional: executable helpers
  references/       # Optional: supporting documentation
  assets/           # Optional: templates, resources

SKILL.md Format

---
name: my-skill-name          # Must match folder name, lowercase + hyphens
description: What it does    # Task-focused sentence (20-600 chars)
license: MIT                 # Required: license type
# Optional semantic fields:
inputs: [source-code]        # What the skill expects
outputs: [test-report]       # What the skill produces
side_effects: [creates-files] # Environment changes
triggers: [user-asks-about-testing] # Activation conditions
complements: [verification-loop]    # Skills that pair well
includes: [skill-a, skill-b]       # Bundle: skills to install together
tier: core                          # Quality tier: core or community
---

# Instructions and content here

Generated Directories (in distributions/, managed by refresh script)

  • distributions/collections/example-skills.txt — skill path list
  • distributions/collections/core-skills.txt / community-skills.txt — tier lists
  • distributions/skills-registry.json — machine-readable skill metadata (all frontmatter + resources)
  • distributions/skills-lock.json — lockfile with SHA-256 hashes per skill
  • distributions/direct/example/ — direct link directory
  • distributions/codex/skills / distributions/claude/skills — agent-specific bundles
  • distributions/extensions/gemini/*/skills — Gemini CLI extensions

These generated artifacts remain tracked until the coordinated installation migration is accepted; include refreshed outputs in PRs that change skills. CI validates that generated files are up-to-date. They are not build inputs: scripts/skills_install.py generates and validates a complete staging directory from canonical source, then atomically activates the verified release. The installer supports clean checkouts, failed-update preservation, guarded rollback, and standalone release archives. See docs/installation.md.

Release Version

.claude-plugin/plugin.json is the authored catalog version. The release helper updates it; the generators derive marketplace and Gemini descriptor versions from it. Never edit a generated descriptor as a release input.

Key Guidelines

  • Skill name in frontmatter must exactly match the directory name
  • All skills must have a license field (MIT for open skills)
  • Skills are organized into category subdirectories; new skills must go in the appropriate category
  • Only redistributable (open-licensed) skills belong here; Anthropic's document skills (docx, pdf, pptx, xlsx) are all-rights-reserved and must not be re-added
  • Repo-wide tests live in tests/ (python3 -m pytest); CI runs them
  • Update docs/THIRD_PARTY_NOTICES.md when adding external assets
  • CI includes secret detection for patterns like sk-, ghp_, AKIA in new files

GitHub MCP Usage Policy

IMPORTANT: This repo exists locally. ALWAYS use local filesystem tools (Read, Grep, Glob) to read files — NEVER use mcp__github__get_file_contents or similar GitHub MCP tools for content that exists on disk. GitHub MCP tools trigger API content filtering errors on repos with security-domain content (security skills contain terms like STRIDE, XSS, SQL injection that trip the output filter).

Use GitHub MCP tools ONLY for operations that require the GitHub API:

  • Setting repo description/topics (update_pull_request, etc.)
  • Creating/managing issues and PRs
  • Managing labels, releases, branches

For these API-only operations, always set minimal_output: true where the parameter is available.

<!-- ORGANVM:AUTO:START -->

System Context (auto-generated — do not edit)

Organ: ORGAN-IV (Orchestration) | Tier: standard | Status: GRADUATED Org: organvm-iv-taxis | Repo: a-i--skills

Edges

  • No inter-repo edges declared in seed.yaml

Siblings in Orchestration

orchestration-start-here, petasum-super-petasum, universal-node-network, .github, agentic-titan, agent--claude-smith, tool-interaction-design, system-governance-framework, reverse-engine-recursive-run, collective-persona-operations, contrib--adenhq-hive, contrib--ipqwery-ipapi-py, contrib--primeinc-github-stars, contrib--temporal-sdk-python, contrib--dbt-mcp ... and 6 more

Governance

  • Standard ORGANVM governance applies

Last synced: 2026-05-23T00:26:31Z

Active Handoff Protocol

If .conductor/active-handoff.md exists, READ IT FIRST before doing any work. It contains constraints, locked files, conventions, and completed work from the originating agent. You MUST honor all constraints listed there.

If the handoff says "CROSS-VERIFICATION REQUIRED", your self-assessment will NOT be trusted. A different agent will verify your output against these constraints.

Session Review Protocol

At the end of each session that produces or modifies files:

  1. Run organvm session review --latest to get a session summary
  2. Check for unimplemented plans: organvm session plans --project .
  3. Export significant sessions: organvm session export <id> --slug <slug>
  4. Run organvm prompts distill --dry-run to detect uncovered operational patterns

Transcripts are on-demand (never committed):

  • organvm session transcript <id> — conversation summary
  • organvm session transcript <id> --unabridged — full audit trail
  • organvm session prompts <id> — human prompts only

System Library

Plans: 269 indexed | Chains: 5 available | SOPs: 8 active Discover: organvm plans search <query> | organvm chains list | organvm sop lifecycle Library: /Users/4jp/Code/organvm/praxis-perpetua/library

Active Directives

ScopePhaseNameDescription
systemanyatomic-clockThe Atomic Clock
systemanyexecution-sequenceExecution Sequence
systemanymulti-agent-dispatchMulti-Agent Dispatch
systemanysession-handoff-avalancheSession Handoff Avalanche
systemanysystem-loopsSystem Loops
systemanyprompting-standardsPrompting Standards
systemanybackground-task-resiliencebackground-task-resilience
systemanycontext-window-conservationcontext-window-conservation
systemanysession-self-critiquesession-self-critique
systemanythe-descent-protocolthe-descent-protocol
systemanythe-membrane-protocolthe-membrane-protocol
systemanytheory-to-concrete-gatetheory-to-concrete-gate
systemanytriangulation-protocoltriangulation-protocol
unknownanySOP-001_REPOSITORY_SEEDINGSOP-001: Repository Seeding Procedure
unknownanySOP-002_WORKSPACE_AUDITSOP-002: Comprehensive Workspace Audit Procedure
unknownanySOP-003_GOVERNANCE_PROMOTIONSOP-003: Governance Promotion Procedure
unknownanySOP-004_SEED_YAML_VALIDATIONSOP-004: Seed.yaml Validation Procedure
unknownanySOP-005_ORGAN_CLASSIFICATIONSOP-005: Organ Classification Procedure
unknownanySOP-006_PHASE_TRANSITIONSOP-006: Phase Transition Procedure
unknownanySOP-007_CLAUDE_MD_GENERATIONSOP-007: CLAUDE.md Generation Procedure
unknownanySOP-008_DEPENDENCY_MAPPINGSOP-008: Dependency Mapping Procedure
unknownanySOP-009_IRF_ASSIGNMENTSOP-009: IRF Assignment Procedure
unknownanySOP-010_MULTI_REPO_ORCHESTRATIONSOP-010: Multi-Repo Orchestration Procedure

Linked skills: SOP-TRIADIC-REVIEW-PROTOCOL, cicd-resilience-and-recovery, continuous-learning-agent, evaluation-to-growth, genesis-dna, multi-agent-workforce-planner, promotion-and-state-transitions, quality-gate-baseline-calibration, repo-onboarding-and-habitat-creation, session-self-critique, structural-integrity-audit, the-membrane-protocol, triple-reference

Prompting (Anthropic): context 200K tokens, format: XML tags, thinking: extended thinking (budget_tokens)

Atomization Pipeline

Run organvm atoms pipeline --write && organvm atoms fanout --write to generate task queue.

System Density (auto-generated)

AMMOI: 25% | Edges: 0 | Tensions: 0 | Clusters: 0 | Adv: 27 | Events(24h): 37975 Structure: 8 organs / 148 repos / 1654 components (depth 17) | Inference: 0% | Organs: META-ORGANVM:63%, ORGAN-I:53%, ORGAN-II:48%, ORGAN-III:54% +5 more Last pulse: 2026-05-23T00:26:28 | Δ24h: n/a | Δ7d: n/a

Dialect Identity (Trivium)

Dialect: GOVERNANCE_LOGIC | Classical Parallel: Rhetoric | Translation Role: The Meta-Logic — governance rules ARE propositions

Strongest translations: I (formal), V (structural), META (structural)

Scan: organvm trivium scan IV <OTHER> | Matrix: organvm trivium matrix | Synthesize: organvm trivium synthesize

Logos Documentation Layer

Status: ACTIVE | Symmetry: 0.5 (DREAM)

Nature demands a documentation counterpart. This formation maintains its narrative record in docs/logos/.

The Tetradic Counterpart

  • Telos (Idealized Form) — The dream and theoretical grounding.
  • Pragma (Concrete State) — The honest account of what exists.
  • Praxis (Remediation Plan) — The attack vectors for evolution.
  • Receptio (Reception) — The account of the constructed polis.

Alchemical I/O

  • Source & Transmutation — Narrative of inputs, process, and returns.

Compliance: Record exists without implementation.

<!-- ORGANVM:AUTO:END -->

⚡ Conductor OS Integration

This repository is a managed component of the ORGANVM meta-workspace.

  • Orchestration: Use conductor patch for system status and work queue.
  • Lifecycle: Follow the FRAME -> SHAPE -> BUILD -> PROVE workflow.
  • Governance: Promotions are managed via conductor wip promote.
  • Intelligence: Conductor MCP tools are available for routing and mission synthesis.

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.