Maintains and optimizes Claude agent definitions through automated quality diagnosis, improvement suggestions, and autonomous evolution protocols.
Agent Quality Manager (agent-brushup) is a comprehensive toolkit designed to manage and evolve the specialized roles stored in your Claude agent directory. It provides a structured pipeline to scan for anti-patterns—such as hardcoded values, missing frontmatter, or vague instructions—and generates high-quality improvements based on industry-standard best practices. Whether you are scaffolding new agents with self-evolution protocols or pruning unused definitions based on telemetry data, this skill ensures your AI development environment remains lean, efficient, and aligned with the latest prompt engineering strategies.
主要功能
01Interactive improvement suggestions with diff-based application and user approval
02Automated quality scoring and anti-pattern detection for agent definitions
03Upstream synchronization with the agency-agents best-practice catalog
04Telemetry-based pruning to identify and remove unused or low-quality agents
051 GitHub stars
06Smart scaffolding for new agents with built-in Self-Evolution Protocols
使用场景
01Managing a high volume of custom agents by identifying and removing redundant or obsolete roles
02Automatically upgrading legacy agent prompts to include modern JIT file reference strategies
03Standardizing team-wide AI agent definitions to prevent instruction drift and technical debt
