Skill UI

A systematic workflow designed to significantly improve the reliability and performance of existing AI agents. It covers detailed performance analysis, including gathering metrics (e.g., task success rate, latency), classifying failure modes, and advanced prompt engineering techniques (like Chain-of-Thought and Constitutional AI). The process culminates in structured A/B testing and controlled deployment, ensuring measurable, robust improvements.

Provides comprehensive guidance on building robust evaluation frameworks for complex agent systems. It addresses the inherent challenges of non-determinism and multiple valid paths, enabling systematic performance testing, validating context engineering choices, and measuring continuous improvements across multiple quality dimensions (e.g., factual accuracy, tool efficiency).

Automates iterative use of the plugin-dev skill-reviewer agent to diagnose and fix Claude Code SKILL.md issues, driving repeated review/fix cycles until standards are met before emitting the completion marker.

This skill provides a systematic methodology for iteratively improving complex skill documentation. By utilizing a reviewer agent, it identifies and categorizes issues (Critical, Major, Minor) in a structured loop. The process ensures that the skill meets high standards of quality, reliability, and functional completeness before deployment, moving beyond simple manual edits.

This skill reviews code differences (diffs) to identify opportunities for improvement in clarity, reusability, efficiency, and overall quality. It employs multiple specialized sub-agents (e.g., Code Reuse, Efficiency, Quality) to audit changes. It can provide high-confidence, behavior-preserving fixes when requested, ensuring the code adheres to best practices without introducing regressions. Best used when refactoring, cleaning up, or auditing a specific code scope.

This framework provides a rigorous, multi-dimensional guide for conducting code reviews. It evaluates five critical axes: correctness, readability, architecture, security, and performance. Use this process before merging any code change, especially when reviewing PRs, refactoring, or evaluating code produced by AI agents, to ensure continuous improvement and maintain high code health.

Provides patterns and techniques for building robust evaluation loops for AI agents. This skill guides the implementation of self-critique, evaluator-optimizer pipelines, and test-driven refinement workflows. Use it to significantly enhance the accuracy and quality of complex outputs like code, reports, and analyses by implementing structured, iterative improvement cycles (Generate → Evaluate → Critique → Refine).

Evolver is a comprehensive self-evolution engine designed for autonomous AI agents. It analyzes complex runtime history, identifies performance bottlenecks, and autonomously generates and applies improvements. Communication is architecturally confined to a local Proxy Mailbox, ensuring secure and reliable interaction with the EvoMap Hub for asset management and task acquisition. Ideal for developing highly adaptive, continuously improving intelligent systems.

This framework enforces high output quality, intellectual honesty, and continuous improvement across all AppSec agents. It mandates structured processes, including self-critique loops, anti-rationalization guards, and self-reflection quality gates (≥8 scoring). Applicable during code review, SAST, SCA, and threat modeling, it ensures rigorous and verifiable security analysis.

This skill provides a direct shell call to capture the current Git diff (staged, working-tree, or specific file list) as raw text. It is designed for workflows requiring repository diffs without the overhead of an LLM agent loop, offering significant performance improvements over traditional sub-agent calls.

SkillOpt-Sleep provides a mechanism for autonomous agent self-improvement. By simulating a 'sleep' or 'dream' cycle, it harvests past interaction sessions, mines recurring tasks, and replays them through the designated backend. It consolidates learned knowledge into persistent memory and skills, ensuring that only strictly validated improvements are adopted via a held-out gate, simulating long-term competence gain from short-term experience. Ideal for scheduled, background optimization.

Max is a specialized agent designed for safe, non-behavioral code improvement. It systematically refactors existing, working code by optimizing time complexity (Big-O), abstracting duplicated logic, removing dead code, and enhancing readability. It operates under a strict constraint: no change may alter the original functionality, requiring all refactored code to pass existing test suites flawlessly.