
ffmpeg-skill
#58A local video editing toolkit for coding agents with 42 FFmpeg-based tools and an MCP server. It runs offline with no API keys.


Plannotator
#53A local browser tool to annotate your coding agent's plans and review its changes, then send the feedback back to the agent.

Google's plugin where one agent drafts the design and fresh-context reviewers check it before you approve the spec.

Scans your codebase and recommends hooks, skills, MCP servers, subagents and slash commands. Read-only.

Code Review
#50Four agents review a pull request and only report issues they score 80 or more out of 100 for confidence.


Warns on about 25 risky code patterns and reviews Claude's diffs and your commits. Reviews run on your own Claude plan.

Superpowers
#50Teaches Claude brainstorming, test-first development, systematic debugging and subagent-driven work with code review.


OpenAI's open-source tool that finds, checks and patches vulnerabilities in your code, as a Codex plugin, CLI, SDK or GitHub Action.

Skills that steer coding agents toward modern, native browser features instead of heavy JavaScript.



A template that turns one Claude Code session into a game studio: 49 agents for Godot, Unity or Unreal, with your sign-off at every step.

Agent skills for Google Cloud, Firebase, BigQuery, Cloud Run, the Gemini API and more.

Builds a semantic layer over your SQL data so AI agents know what it means, served through MCP, a CLI and a REST API.




claude-mem
#13Gives Claude Code memory across sessions: hooks capture what the agent does and feed the relevant parts back later.

A CLI that lets AI agents use desktop apps through the accessibility tree instead of screenshots, on macOS, Windows and Linux.

NVIDIA's open-source scanner that checks AI agent skills for prompt injection, data exfiltration and supply-chain risks before you install them.
