checkmate
Enforces task completion: turns your goal into pass/fail criteria, runs a worker, judges the output, feeds back what's missing, and loops until every criterion passes. Nothing ships until it's truly done. Trigger: 'checkmate: TASK'
Why use this skill?
Use the Checkmate skill to enforce quality standards in OpenClaw. It automates iterative feedback loops, ensuring tasks meet strict criteria before completion.
Install via CLI (Recommended)
clawhub install openclaw/skills/skills/insipidpoint/checkmateWhat This Skill Does
Checkmate is a high-assurance orchestration skill for OpenClaw that enforces absolute task completion through a rigorous iterative feedback loop. Unlike standard agent interactions that aim for a "best effort" response, Checkmate transforms your goals into concrete pass/fail criteria. It deploys a Python-based execution loop where one LLM acts as the worker to perform the task, while a secondary LLM acts as the judge to audit the results against your defined metrics. The process continues recursively—feeding back gaps and missing components—until every requirement is satisfied. Nothing is finalized or "shipped" until the judge confirms total compliance.
Installation
To integrate Checkmate into your OpenClaw environment, execute the following command in your terminal:
clawhub install openclaw/skills/skills/insipidpoint/checkmate
Ensure that you have Python 3 installed and the openclaw CLI configured in your system PATH. The tool leverages existing OAuth sessions from your gateway, meaning no additional API keys are required for operation.
Use Cases
Checkmate is designed for high-stakes scenarios where accuracy is non-negotiable. It is ideal for:
- Software Development: Generating code that must pass specific unit tests or architectural specs.
- Technical Documentation: Producing reports or whitepapers that require adherence to strict style guides, citation requirements, or content coverage.
- Data Auditing: Verifying that research tasks have explored all necessary data points and exhausted all relevant sources.
- Complex Workflows: Any project that usually requires the human user to provide multiple "try again" prompts until the output matches expectations.
Example Prompts
- "checkmate: Write a Python script to scrape the provided URL for stock prices, ensuring it handles 404 errors gracefully and logs output to a CSV."
- "I need a detailed project plan for the Q3 migration; keep iterating until it passes my requirement that all security compliance checklists are addressed."
- "quality loop: Research the history of the company, judge and retry until the summary includes at least three primary sources and covers all executive leadership changes."
Tips & Limitations
- Interactive vs. Batch: In default interactive mode, you must approve checkpoints. Use the
--no-interactiveflag only for trusted, low-risk automation tasks. - Privilege Awareness: Workers spawned by Checkmate inherit full agent privileges, including
exec,web_search, and access to your OAuth-linked accounts. Only execute tasks from sources you trust. - Persistence: User input provided during a checkpoint is written to
user-input.md. Avoid pasting sensitive system tokens or third-party untrusted data into this interface. - Performance: Because this skill forces iterative improvement, it consumes more token credits than standard agent tasks. Use it when quality is the priority over speed.
Metadata
Not sure this is the right skill?
Describe what you want to build — we'll match you to the best skill from 16,000+ options.
Find the right skillPaste this into your clawhub.json to enable this plugin.
{
"plugins": {
"official-insipidpoint-checkmate": {
"enabled": true,
"auto_update": true
}
}
}Tags(AI)
Flags: network-access, file-write, file-read, external-api, code-execution