How to Install & Use
Your Governance Skill
No prior experience required. If you can open a terminal, you can deploy runtime governance in under 10 minutes.
Unzip Your Download
Each purchase downloads as a .zip file containing all governance files, tool scripts, and documentation.
Right-click the .zip file and extract it. You'll see a folder containing:
your-skill-name/
├── CLAUDE.md ← Skill manifest
├── GETTING-STARTED.md ← This guide
├── README.md ← Operations manual
├── policy.yaml ← Governance rules
├── middleware.ts ← Runtime code
├── architecture.mmd ← Architecture diagram
├── financial-model.csv ← ROI metrics
└── tools/ ← Reusable scripts
├── tool-1.sh
├── tool-2.sh
├── tool-3.sh
└── tool-4.shCopy Into Your Project
Create a skills directory in your project and copy the governance skill into it.
Open your terminal and run these two commands:
# Create the skills directory mkdir -p .claude/skills/ # Copy the skill folder into your project cp -r /path/to/your-skill-name/ .claude/skills/your-skill-name/
Reference It in CLAUDE.md
Tell your AI agent to load and follow the governance rules.
Open (or create) CLAUDE.md in your project root and add:
## Governance Skills Load and follow the governance rules in: - .claude/skills/your-skill-name/CLAUDE.md - .claude/skills/your-skill-name/policy.yaml When any trigger condition is met, execute the corresponding tool script in the tools/ directory.
Verify It Works
Start a new agent session and confirm the skill is loaded.
Ask your agent:
"What governance skills are loaded? List the trigger conditions and escalation levels."
Agent-Specific Setup
These skills work with every major AI coding agent. Here's where to put the files for each one:
Understanding policy.yaml
The policy file defines when the governance system activates and what it does. Here's what the thresholds mean in plain English:
| If This Happens... | The Agent Should... |
|---|---|
| Context window 50% full | Log a warning |
| Context window 65% full | Save state and reset |
| Same file patched 5+ times | Stop and ask you |
| Spending >$15/hr on tokens | Kill process immediately |
You can customize these. Edit the numbers in policy.yaml to match your team's risk tolerance.
Combining Multiple Skills
For maximum protection, deploy one skill from each governance layer:
Troubleshooting
Agent doesn't mention the skill
Make sure your root CLAUDE.md or rules file references the skill path. Start a new session after adding the reference.
Agent ignores the governance rules
Rules are loaded at session start. Restart the session. If using Cursor, close and reopen the editor.
Tool scripts won't run
On Mac/Linux, run: chmod +x tools/*.sh — On Windows, use Git Bash or WSL.
YAML syntax error in policy.yaml
YAML requires spaces, not tabs. Open in VS Code which highlights syntax errors automatically.
Agent runs tools but they produce errors
Make sure you're running from the project root, not from inside the skill folder. The tools use relative paths.
Ready to deploy?
Choose your governance system and start protecting your codebase in under 10 minutes.