A rigorous quality assurance protocol that forces AI agents to verify every file modification twice through independent tests or runtime checks before completion.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install double-check
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 double-check 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 Double-Check Skill?
The Double-Check skill is a specialized reliability protocol designed for Openclaw Skills to eliminate regression errors and accidental breakages during the development process. It mandates that any file modification—including code updates, feature removals, or configuration changes—must be followed by two consecutive, successful validation cycles. By ensuring that an AI agent cannot mark a task as complete without passing these independent checks, it provides a safety net for automated refactoring and bug fixing.
This skill is built on the principle that file quality is a deliberate outcome of systematic verification rather than a side effect. It effectively bridges the gap between AI-generated code and production-ready stability by forcing a loop of modification, verification, and correction until a stable state is confirmed twice.
Double-Check Skill Use Cases
- Bug Resolution: Automatically running unit tests twice after a fix to ensure no intermittent failures exist.
- Feature Removal: Verifying that deleting obsolete code doesn't leave dangling references or break the build.
- Refactoring: Using compiler checks or linters to ensure structural changes haven't introduced syntax or type errors.
- Dependency Updates: Confirming that new library versions don't crash the application during runtime or build phases.
- Documentation & Config: Validating that markdown builds or schema checks pass after content updates.
How Double-Check Skill Works
- Intent Recognition: The AI detects a request to modify, add, or delete files and activates the protocol with a ✅✅ indicator.
- Execution: The agent performs the requested file changes across the codebase.
- Validation Selection: The agent identifies the appropriate verification method, such as
npm test,pytest, orcargo checkbased on the project environment. - First Verification: The verification command is executed; if it fails, the agent fixes the error and restarts the count.
- Second Verification: Upon the first success, the agent runs the exact same command again (often with
--no-cache) to ensure the result is consistent and independent. - Final Completion: The agent reports the double-success status and concludes the task only after two consecutive passes.
Double-Check Skill Setup
To integrate this protocol into your environment using Openclaw Skills, ensure your agent has access to the local shell and project-specific test runners. No complex installation is required beyond adding the skill definition to your agent's profile.
# Example of a verification command the skill might trigger
npm test -- --no-cache
# Or for Python projects
pytest --cache-clear
Double-Check Skill Data Schema & Taxonomy
The skill operates by tracking the state of verification cycles. It does not generate standalone metadata files but follows a strict internal taxonomy for reporting:
| Attribute | Description |
|---|---|
| Activation Mark | Every response starts with ✅✅ to signal protocol adherence. |
| Verification Count | A counter (0-2) tracking consecutive successful runs. |
| Exit Codes | Only an exit code of 0 is treated as a success for verification steps. |
| Error Logs | Detailed capture of stderr/stdout to inform the 'Correction' phase. |
| Cache Policy | Prefers no-cache flags to ensure data integrity between runs. |
Double-Check Skill Advanced Features
- Cache Busting: Automatically appends flags like
--no-cacheor--cache-clearto ensure verifications are truly independent. - Bilingual Support: Fully documented for both English and Chinese speaking development teams.
- Framework Integration: Seamlessly works alongside the answer-framework to provide structured reasoning before the verification loop begins.
- Fallback Negotiation: If no standard test suite is found, the agent proactively negotiates a custom runtime check with the user.
SKILL.md
Loading
Related Openclaw Skills
Ninox Integration for AI Agents
A powerful low-code database integration that enables AI agents to manage business data, CRM records, and automated workflows seamlessly.
Onelap Ride Download
A specialized tool for batch downloading cycling activity FIT files from the Onelap platform for local backup and multi-platform analysis.
Deep Reading
A high-fidelity framework for transforming books, papers, and long-form articles into structured, actionable, and networked knowledge.
Akeyless CLI
A secure integration for managing Akeyless secrets, static/dynamic credentials, and gateway routing via the official CLI.
Homestruk Lease Renewal Tracker
A proactive property management skill that automates the 90-day lease renewal lifecycle to maximize tenant retention and optimize rental income.
Find Skills
A discovery and installation engine for extending AI agent capabilities through a modular package ecosystem.
Featured*
