Accountable AI
by @kgapol
Implements governance for AI teams by defining roles, decision authority, procedures, and delegation to ensure accountability and traceable actions.
clawhub install accountable-aiπ About This Skill
Accountable AI β Governance Framework for AI Agent Teams
Use this skill to implement accountability structures, decision trails, and responsibility chains across your AI agent team.
What This Skill Does
Establishes a governance framework that makes every AI agent in your team accountable for:
When to Use
Core Files
This skill installs three files to your agent workspace:
GOVERNANCE.md β The accountability charter. Defines roles, decision authority, and responsibility chains. Each agent reads this on startup to know what they can do independently and what requires escalation.
PROCEDURES.md β Repeatable checklists for operations that must not be improvised. Config changes, agent onboarding, infrastructure tasks. When an agent encounters a known procedure type, it runs the checklist β not its own interpretation.
DELEGATION.md β The delegation protocol. How work moves between agents. Who owns what. How handoffs are documented. Prevents the "I thought someone else was handling it" failure mode.
How It Works
1. Install the files to your agent workspace 2. Reference them in your AGENTS.md so agents load them each session 3. Agents follow the governance rules automatically when making decisions 4. Procedures prevent improvisation on sensitive operations 5. Delegation logs create an audit trail of who did what
Installation
chmod +x install.sh && ./install.sh
Or manually copy GOVERNANCE.md, PROCEDURES.md, and DELEGATION.md to your agent workspace.
After Installing
Add to your AGENTS.md boot sequence:
3. Read GOVERNANCE.md β your decision authority and escalation rules
4. Read PROCEDURES.md β before any config change or infrastructure task
Part of Trust AI Suite
Accountable AI is the foundation layer of the Trust AI Suite:
β Full suite: Trust AI Stack on ClawMart