A professional governance engine that enforces rigorous evidence-based protocols and safety gates for AI coding agents.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install yes-md-ja
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install yes-md-ja using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is YES.md - AI Governance Engine?
YES.md is a high-performance governance framework designed to transform AI coding agents into professional-grade engineers. While other Openclaw Skills might push an agent to keep trying, YES.md ensures that every attempt is safe, verified, and backed by concrete evidence. It replaces AI intuition and guessing with a structured architecture of verification, ensuring that agents act as partners who deliver validated results rather than just reporting effort.
The skill is built upon three pillars: Safety Gates to prevent system regressions, Evidence Rules to mandate data-driven diagnostics, and Ripple Awareness to ensure that every code change is analyzed for its side effects. By integrating this into your workflow, you eliminate the common patterns of AI laziness, such as premature hand-offs to the user or unverified code proposals.
YES.md - AI Governance Engine Use Cases
- High-stakes debugging where evidence must be proven through logs and CLI tools before proposing a fix.
- Secure configuration management where automatic backups must be created before any file modification.
- Complex system migrations requiring impact analysis and dependency checking via grep and lsof.
- Professional deployment workflows where every change must be verified with automated tests or curl commands.
- Preventing AI circular reasoning by enforcing a pivot in approach after two failed attempts.
How YES.md - AI Governance Engine Works
- Monitoring: The skill monitors agent behavior for signs of guessing, lazy questioning, or repetitive failure patterns.
- Enforcement: It triggers mandatory 'Safety Gates' requiring the agent to perform backups (e.g., cp file file.bak) and impact assessments.
- Evidence Gathering: Agents are forced to use available tools like Bash, Read, and WebSearch to prove their hypotheses before speaking.
- Verification: After any modification, the agent must execute specific verification steps to demonstrate the fix is functional.
- Closing: The process concludes with a Bug-Close Protocol that records the root cause and lessons learned for future Openclaw Skills sessions.
YES.md - AI Governance Engine Setup
To deploy this governance engine within your environment, ensure your agent has access to the core SKILL.md file and the following permissions:
# Ensure your AI agent has execute permissions for the following toolsets:
# 1. Bash / CLI for system verification
# 2. File System access for creating .bak files
# 3. WebSearch for external documentation lookups
Integrate the protocol by referencing the YES.md logic in your system instructions to enable professional-grade behavior across all Openclaw Skills.
YES.md - AI Governance Engine Data Schema & Taxonomy
The skill organizes its operational data and metadata into a strict taxonomy of verification:
| Component | Organization Method | Metadata Captured |
|---|---|---|
| Backups | File duplication with suffixes | Original path, timestamp, reason |
| Evidence | Command-line output capture | Data source, time range, sample size |
| Debug Audit | 5-Step structured log | Error context, assumption testing, hypothesis |
| Handover | Structured Markdown documentation | Verified facts, excluded causes, next steps |
YES.md - AI Governance Engine Advanced Features
- 5-Step Debug Audit: A mandatory escalation protocol that triggers after three failures to force a deep-dive into error context.
- Automated Safety Gates: Non-negotiable requirements for backups and impact analysis before any system-level file edit.
- Structured Handoff Logic: Ensures that if an agent must stop, it provides a professional-grade technical briefing for the human developer.
- Self-Correction Mechanism: A built-in detector for 'lazy phrases' like 'I think' or 'please check,' forcing the agent to find the answer itself.
- Ripple Check: An automated post-fix scan to ensure the change hasn't broken upstream or downstream dependencies within the codebase.
SKILL.md
Loading
Related Openclaw Skills
GIS Coordinate Transform
A high-performance command-line utility for transforming GeoJSON coordinates between WGS84 and Web Mercator projections.
Quick Translation Skill
A high-performance AI translation tool supporting over 100 languages with advanced terminology recognition and style customization.
Professional Knowledge Extraction Skill
A sophisticated knowledge extraction engine that transforms unstructured professional text into multi-layered, interactive knowledge graphs.
HTTP Security Audit
A comprehensive security scanner that audits HTTP response headers and information leaks to help harden web server configurations.
Brave Search Custom
A specialized skill enabling AI agents to perform privacy-conscious web searches using the Brave Search API.
Chef Dialog Recipe
A professional-grade AI culinary agent that generates, audits, and optimizes recipes through a multi-stage chef dialogue workflow.
Featured*
