compound-writing — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited compound-writing (Plugin) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Extend Claude Code with specialized workflows, automatic formatting, and better defaults.
Skills give Claude domain expertise—brainstorming methods, documentation generation, book writing pipelines. Hooks automate repetitive tasks—format code after edits, summarize changes at session end. Templates configure how Claude works with you.
Works with both Claude Code (CLI) and Claude.ai (web/mobile/desktop).
# Clone the toolkit
git clone https://github.com/robertguss/claude-code-toolkit.git
# Use a skill in Claude Code
# Add to your project's CLAUDE.md:
echo "When brainstorming, read /path/to/claude-code-toolkit/skills/brainstorm/SKILL.md" >> CLAUDE.md
# Or package for Claude.ai
python build.py brainstorm
# Upload dist/brainstorm.skill to Claude.ai → Settings → SkillsPackaged workflows that Claude follows when invoked. Use them with /skillname or reference in CLAUDE.md.
| Skill | Description |
|---|---|
| brainstorm | Multi-session ideation partner with method catalog |
| code-documenter | Intelligent documentation generation with health tracking |
| handoff | Session continuity documents for picking up where you left off |
Book & Writing:
from idea to chapter architecture
Shell scripts that run automatically at specific Claude Code events.
| Hook | Event | Description |
|---|---|---|
| auto-format | PostToolUse | Formats files after edits (ruff, goimports, prettier) |
| change-summary | Stop | TypeScript checking + session change summary |
| compaction | PreCompact | Injects preservation priorities for better context compaction |
Document templates for configuring how Claude works with you.
| Template | Purpose |
|---|---|
| HUMAN.md | Relationship document — helps Claude remember who you are |
| CLAUDE.md | Global instructions for all projects |
| compaction-strategy.md | What to preserve during context compaction |
Claude Code (CLI):
Reference in your project or global CLAUDE.md:
# CLAUDE.md
When brainstorming, read and follow
/path/to/claude-code-toolkit/skills/brainstorm/SKILL.mdClaude.ai (Web/Mobile/Desktop):
python build.py brainstorm
# Upload dist/brainstorm.skill to Claude.ai → Settings → Skills cp hooks/auto-format/auto-format.sh ~/.claude/hooks/
chmod +x ~/.claude/hooks/auto-format.sh~/.claude/settings.json: {
"hooks": {
"PostToolUse": [
{
"matcher": "Edit|MultiEdit|Write",
"hooks": [
{
"type": "command",
"command": "$HOME/.claude/hooks/auto-format.sh"
}
]
}
]
}
}See each hook's README for specific configuration.
# Copy and customize
cp templates/HUMAN.md ~/.claude/YOURNAME.md
cp templates/CLAUDE.md ~/.claude/CLAUDE.md
# Edit to match your preferencesclaude-code-toolkit/
├── skills/ # Packaged workflows
│ ├── brainstorm/
│ ├── code-documenter/
│ ├── handoff/
│ └── ...
├── hooks/ # Automatic event handlers
│ ├── auto-format/
│ ├── change-summary/
│ └── compaction/
├── templates/ # Configuration templates
│ ├── HUMAN.md
│ ├── CLAUDE.md
│ └── compaction-strategy.md
├── docs/ # Documentation site source
└── build.py # Skill packager for Claude.aiThis project uses uv for dependency management and just as a command runner.
# Install dependencies
just install
# Serve docs locally at http://localhost:8000
just docs-serve
# Deploy docs to GitHub Pages
just docs-deploy
# Package a skill
just package brainstorm
# See all commands
justSee CONTRIBUTING.md for guidelines on submitting skills, hooks, or improvements.
MIT License. See LICENSE.md for details.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.