Skill UI

A comprehensive multi-agent feedback loop inspired by Generative Adversarial Networks (GANs), designed to build complex, production-ready applications. It separates generation from evaluation, using a 'ruthlessly strict' evaluator (QA) to drive continuous iteration. This process ensures high design quality, robust functionality, and originality, overcoming the limitations of single-agent self-critique. Ideal for full-stack feature development and complex product specification implementation.

A utility for scaffolding and verifying fixed evaluation corpora (bench suites). It generates deterministic test suites from existing repositories, decoupling the evaluation process from the repo's natural test suite. This ensures consistent fitness scoring when running evolution pipelines across different commits or forks.

This tool simplifies the process of detecting structural drift by combining audit listing, fresh auditing, and trend diffing into a single command. It finds the most recent audit record, runs a current audit against the repository, and compares the two using structural similarity metrics. It alerts when the structural distance falls below a specified threshold, helping maintain code consistency over time.

This skill runs an automated evolution process designed to systematically optimize the seven core policy surfaces of an AI agent harness (e.g., planner, contextBuilder, reviewer). By mutating these policies and scoring the variants in a sandbox, it identifies crucial performance improvements without requiring expensive foundation model retraining. It is ideal for empirical configuration tuning and establishing optimized baselines.

A comprehensive 7-section report providing a deep assessment of a repository's health and readiness level. It calculates key metrics including risk score, agent topology recommendations, MCP surface area, test confidence, and publish readiness. This tool is vital for pre-mint reviews, detecting architectural drift, and enforcing strict CI/CD quality gates.

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 tool executes a GEPA learning cycle, automating the evolution and optimization of policy prompts against a scored task corpus. Instead of relying on manual prompt iteration, it measures performance improvements, identifying the best working 'harness genome.' It is ideal for measuring sustainable improvements in task performance and supports a free, cost-estimating dry-run mode before committing to actual model calls.

Performs a pure, read-only static security scan of a harness's declared MCP surfaces. It analyzes policy, permission, and dependency issues by reading manifest files (e.g., .mcp/servers.json and .harness/claims.json) without executing any underlying tools. Designed for integration into CI/CD pipelines to enforce security policies and ensure compliance.

A powerful tool designed to scaffold sophisticated, custom AI agent harnesses using the metaharness CLI. It supports various templates (e.g., coding, devops, marketing) and AI hosts (e.g., Claude, Codex). Crucially, it operates with maximum safety, defaulting to a dry-run mode and refusing to write to the calling repository root, ensuring controlled and predictable development workflows.

A comprehensive audit worker that integrates Open Infrastructure Architecture (OIA) manifest analysis, threat modeling, and MCP security scanning. It computes a composite worst-case severity score and persists time-stamped audit records in a memory namespace. Primarily designed for scheduled CI/CD pipelines to detect architectural drift and security regressions over time.

This skill executes the metaharness score to provide a comprehensive, multi-dimensional readiness scorecard for a codebase. It assesses critical dimensions like compilation confidence, task coverage, tool safety, and memory usefulness. Use it to evaluate if a repository is prepared for harness adoption before initiating development tools or framework setup.

This tool executes the advanced `metaharness/darwin` security benchmark (ADR-155). It rigorously evaluates the security detection capabilities of AI systems against a controlled ground-truth corpus, providing critical metrics like TPR, FPR, and patch-pass rates. It compares the champion model against multiple baselines, offering empirical data crucial for iterative security improvement and drift detection.