Execute and Judge Loop
Single-task execution with LLM-as-Judge verification in an iterative loop, supporting auto-retry and strict orchestrator role separation
What is it?
Single-task execution with LLM-as-Judge verification in an iterative loop, supporting auto-retry and strict orchestrator role separation Built for use cases involving execute-verify, llm-as-judge, retry-loop, quality-gate, orchestration.
How to use it?
Install this skill in your Claude environment to enhance execute and judge loop capabilities. Once installed, Claude will automatically apply the skill's guidelines when relevant tasks are detected. You can also explicitly invoke it by referencing its name in your prompts.
The full source and documentation is available on GitHub.
Key Features
- Single-task execution with LLM-as-Judge verification in an iterative loop, supporting auto-retry and strict orchestrator role separation
- Seamless integration with Claude's development workflow
- Comprehensive guidelines and best practices for execute and judge loop
Related Skills
More from AI & MLAgent Evaluation Framework
Comprehensive Claude Code agent evaluation framework with multi-dimensional scoring, LLM-as-Judge mode, and research-backed performance variance analysis
Self-Reflection Framework
Iterative self-improvement system with task complexity grading, strict quality gatekeeper role, confidence thresholds, and verification checklists
Multi-Perspective Critique
Multi-perspective review system using Multi-Agent Debate and LLM-as-Judge patterns with 3 specialized judges, debate rounds, and consensus building