Skill UI

This guide provides a comprehensive reference on Agent Skills, a standardized and powerful method for providing structured context, workflows, and tools to Large Language Models (LLMs). It details the anatomy of the SKILL.md file, the three-stage progressive disclosure loading process, and best practices for creating both capability (tool-based) and process (workflow-based) primitives. Use this resource when designing, reviewing, or debugging agent-driven applications to ensure reliable and context-efficient performance.

This skill executes the DeepSWE benchmark, a comprehensive coding-agent evaluation suite. It uses OpenRouter and mini-swe-agent to score various Large Language Models (LLMs) on 113 tasks, providing reproducible results for model comparison, development assessment, and research benchmarking. Use this when a standardized, deep evaluation of LLM coding capabilities is required.

A specialized tool for generating long-form content, including guides, blogs, and tutorials, that maintains a unique, human-like brand voice. It ensures structural integrity and factual accuracy, preventing the generic, boilerplate feel often associated with raw LLM output. Ideal for content teams needing sophisticated, reliable narrative development.

A comprehensive guide covering patterns and architectures for automating code execution using LLMs. It spans simple sequential pipelines (e.g., daily dev scripts) to advanced, RFC-driven multi-agent Directed Acyclic Graph (DAG) orchestration. Ideal for building CI/CD style continuous development pipelines, managing complex state, and automating quality gates without human intervention.

A high-performance utility for executing simple, deterministic code transformations (e.g., removing console logs, converting var to const). This engine leverages a WASM core to perform structural refactoring sub-millisecond, without calling any expensive LLM APIs. Ideal for CI/CD pipelines, automated bulk refactoring, and cost-sensitive development workflows where determinism and speed are paramount.

This tool provides deep inspection and auditing capabilities for GEPA (genetic-evolution prompt-adaptation) genomes by utilizing the internal `@metaharness/darwin/gepa` library. It is crucial for advanced AI development pipelines, allowing users to validate the structural integrity of genomes, render the full system prompt derived from a genome, or classify specific failure modes found in an execution transcript. Ideal for CI/CD gating and debugging complex LLM applications.

This skill automates the Test-Driven Development (TDD) repair loop. Given a failing unit test, it spawns a bounded, headless LLM agent to automatically modify the source code until the test passes. It closes the development loop by treating the test's pass/fail status as the definitive fitness function, ideal for CI failure triage and local TDD workflows.