agentleFS
appuniteappunite/skills/CLAUDE.md

skills

CLAUDE.md1 starsChanged 8 months agoMITView on GitHub

This is a collection of Claude Code skills for Appunite organization. Skills are specialized capabilities that extend Claude Code with domain-specific knowledge, workflows, and tool integrations. Each skill is a self-contained module that can be installed via npx skills add appunite/skills/<skill-name>. Key Conventions: - skill.json: Defines metadata, invocation command, categories, keywords, dependencies - SKILL.md: Instructions Claude reads when skill is invoked. Uses YAML frontmatter for argument definitions - README.md: Installation, usage examples, configuration for end users - references/: Immutable resources…

  • Installs packages

The file

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

8 rules. About 0 are common, 3 are distinctive, and 5 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.

  • 5

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    MANDATORY: Update CHANGELOG.md before committing changes:

    line 12
  2. Very specific:

    Update version in skill.json and document changes in skill's README.md.

    line 171
  3. Distinctive:

    Test thoroughly: Verify across common scenarios and edge cases

    line 178
  4. Very specific:

    Security first: Use _shared/validation.py for any user input processing

    line 179
  5. Very specific:

    Update CHANGELOG.md: Document changes in [Unreleased] section (MANDATORY before commit)

    line 181
  6. Very specific:

    Package: Run ./prepare-skills.sh to validate and create bundles

    line 182
  7. Very specific:

    PR review: All skills require peer review before merging

    line 183
  8. Distinctive:

    Polish diacritics: Always preserve ą, ć, ę, ł, ń, ó, ś, ź, ż in content

    line 191

≈ “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 a collection of Claude Code skills for Appunite organization. Skills are specialized capabilities that extend Claude Code with domain-specific knowledge, workflows, and tool integrations. Each skill is a self-contained module that can be installed via npx skills add appunite/skills/<skill-name>.

Development Commands

Before Every Commit

MANDATORY: Update CHANGELOG.md before committing changes:

# 1. Document your changes in CHANGELOG.md under [Unreleased]
# 2. Use categories: Added, Changed, Fixed, Removed, Deprecated, Security
# 3. Prefix with skill name for multi-skill changes
# 4. Be specific and user-focused

# Example entry:
# ### Added
# - **roam-transcript-sync**: Pre-flight dependency checker

Skill Packaging

# Validate and package all skills for distribution
./prepare-skills.sh

# This creates versioned bundles in bundles/ directory:
# - Validates required files (skill.json, SKILL.md, README.md)
# - Creates .tar.gz and .zip archives
# - Generates SHA256 checksums
# - Excludes node_modules from references/

Python Skills Testing

Some skills (e.g., roam-transcript-sync) include Python scripts:

# Install Python dependencies
pip install -r roam-transcript-sync/requirements.txt

# Run validation tests
cd _shared && python -m pytest test_validation.py -v

# Run with coverage
python -m pytest test_validation.py --cov=validation --cov-report=html

Architecture

Skill Structure

Each skill follows a standardized directory structure:

skill-name/
├── skill.json          # Metadata: name, version, description, invocation
├── SKILL.md            # Detailed instructions for Claude (YAML frontmatter + markdown)
├── README.md           # User-facing documentation
├── references/         # Supporting files (templates, examples, data)
└── scripts/            # Optional: Python/shell scripts for automation

Key Conventions:

  • skill.json: Defines metadata, invocation command, categories, keywords, dependencies
  • SKILL.md: Instructions Claude reads when skill is invoked. Uses YAML frontmatter for argument definitions
  • README.md: Installation, usage examples, configuration for end users
  • references/: Immutable resources that Claude can reference (never modified during skill execution)

Shared Validation Module

_shared/validation.py provides security controls for all skills that process user input:

Purpose: Centralized defense against path traversal, YAML injection, SSRF, XSS, command injection, race conditions, and DoS attacks.

Used By: meeting-memo, roam-transcript-sync, and any skill handling file operations or external input.

Import Pattern (from any skill's Python script):

import sys
from pathlib import Path

# Add shared module to path
skills_path = Path(__file__).parent.parent
sys.path.insert(0, str(skills_path / "_shared"))

from validation import (
    sanitize_participant_name,
    sanitize_meeting_title,
    validate_notion_url,
    safe_yaml_frontmatter,
    atomic_append_to_file,
    ValidationError,
    SecurityError
)

Critical Functions:

  • sanitize_filename(), sanitize_participant_name(): Prevent path traversal
  • safe_yaml_frontmatter(): Prevent YAML deserialization attacks
  • validate_notion_url(): Prevent SSRF
  • atomic_append_to_file(): Prevent race conditions
  • sanitize_wikilink(): Prevent XSS in markdown

Skill Distribution

Bundles: The bundles/ directory (gitignored) contains packaged versions created by prepare-skills.sh. Each bundle is versioned (e.g., appunite-brand-v1.2.0.tar.gz) and includes checksums for integrity verification.

Installation Methods:

  1. Recommended: npx skills add appunite/skills/<skill-name> (downloads from GitHub)
  2. Manual: Clone repo and symlink to ~/.claude/skills/ or copy directly

Multi-Skill Integration

Some skills depend on others:

  • roam-transcript-sync → automatically invokes meeting-memo for transcript processing
  • Skills declare dependencies in skill.json via integrations.skills array

Python Script Patterns

Skills with Python automation (like roam-transcript-sync) follow these patterns:

Directory Structure:

skill-name/
├── scripts/
│   ├── preflight.py       # Pre-flight checks and setup
│   ├── validation.py      # Symlink to _shared/validation.py
│   ├── failover_utils.py  # Error recovery mechanisms
│   └── sync_state.py      # State management
└── requirements.txt       # Python dependencies

Error Handling: Scripts implement defense-in-depth with:

  • Input validation before processing
  • Atomic file operations with locking
  • Graceful degradation on partial failures
  • Comprehensive logging of security events

Obsidian Integration

Several skills (meeting-memo, roam-transcript-sync) integrate with Obsidian vaults:

Vault Structure Assumptions:

  • Daily notes: DailyNotes/YYYY/MM/YYYY-MM-DD.md
  • Meetings: 2-Areas/Meetings/<person-name>/YYYY-MM-DD topic.md
  • PARA method organization (Projects, Areas, Resources, Archives)

Wikilink Conventions:

  • People: [[Person Name]]
  • Projects: [[Project Name]]
  • Meetings: [[YYYY-MM-DD topic|short description]]

Frontmatter Standards:

---
title: "Meeting title"
person_name: "Person Name"
date: "YYYY-MM-DD"
tags: [meeting, tag1, tag2]
---

Semantic Versioning

Skills follow SEMVER:

  • Major (X.0.0): Breaking changes to skill behavior or invocation
  • Minor (1.X.0): New features, backward compatible
  • Patch (1.0.X): Bug fixes and improvements

Update version in skill.json and document changes in skill's README.md.

Contributing New Skills

  1. Create skill directory with required files (skill.json, SKILL.md, README.md)
  2. Define clear scope: Single responsibility, well-defined purpose
  3. Write instructions: SKILL.md should be unambiguous, with concrete examples
  4. Test thoroughly: Verify across common scenarios and edge cases
  5. Security first: Use _shared/validation.py for any user input processing
  6. Document: Clear README with installation, usage examples, configuration
  7. Update CHANGELOG.md: Document changes in [Unreleased] section (MANDATORY before commit)
  8. Package: Run ./prepare-skills.sh to validate and create bundles
  9. PR review: All skills require peer review before merging

File Patterns to Preserve

  • YAML frontmatter format in SKILL.md (args definitions)
  • skill.json schema (name, description, version, invocation, categories, keywords)
  • Shared validation import pattern for Python scripts
  • Bundle naming convention: <skill-name>-v<version>.tar.gz
  • Polish diacritics: Always preserve ą, ć, ę, ł, ń, ó, ś, ź, ż in content

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.