A curated collection of Claude Code plugins focused on developer workflows, code generation, and productivity.
Agent
Executes beads CLI commands to create issues from a decomposition plan. Use this agent when you have an approved decomposition and need to create the actual beads issues with proper ordering and dependencies.
Agent
Reviews code quality after spec compliance passes. Checks for style, security, error handling, and maintainability issues.
Agent
Reviews task implementations for spec compliance. Use after a subagent completes a task to verify all requirements were met.
Agent
Do NOT use mcpcapacities tools at any point in this workflow. Use $CAP (the cap CLI) for ALL Capacities operations: cap types, cap validate, cap create, cap search, cap link, cap get.
Agent
Use this agent to extract content from research papers (PDFs), YouTube videos, and web articles for the Feynman learning process. Handles PDF downloads, video transcripts, and web content fetching.
Agent
Scans committed code diffs for refactoring opportunities across the codebase. Runs three passes: semantic similarity (Fowler structural patterns + GoF), structural analysis (code smells + SOLID/DRY violations), and language idiom checks. Returns a candidate list for the validator agent to confirm.
Agent
Correlates raw refactoring candidates from scanner shards into architectural root-cause findings. Reads candidates/.yaml from a refactor-scan working directory and writes a human-reviewable findings.md. Does not file beads issues; that is the validator's job.
Agent
Validates refactoring candidates from the scanner agent. Reads actual source files to confirm similarity, checks test coverage, deduplicates against existing beads issues, and creates new issues with TDD-first refactoring plans.
Agent
Use this agent proactively when users want to visualize, diagram, or create visual representations of conversations, code architecture, or concepts. Generates Excalidraw sketch notes with customizable hand-drawn styling.