Back to directory

Agentic Plugin Marketplace

A marketplace of agentic building blocks, including subagents, skills and slash commands, built from one source and installable natively into Claude Code, Codex CLI, Cursor, OpenCode, GitHub Copilot and other agent harnesses.

Agentic Plugin Marketplace

The marketplace ships 94 plugins built from a single Markdown source of truth, each isolated and composable so that installing one plugin loads only its own agents, commands and skills rather than the whole marketplace. Its contents include 202 domain-expert subagents covering areas such as architecture, languages, infrastructure, security, data, machine learning, documentation and business tasks, 183 modular skills that load only once activated, 105 slash commands for scaffolding, security scans, test generation and infrastructure setup, and 16 orchestrators that coordinate multi-agent workflows such as full-stack development, security review or incident response.

Claude Code installs plugins directly through its own plugin marketplace command; Codex CLI and Cursor read the same source through committed registries; Antigravity, OpenCode and Pi install through a clone-and-generate step. Skills alone can also be installed into any of these agents with separate installer commands, without pulling in the accompanying agents or commands.

It is written in Python and licensed under the MIT license.

More in Agent skills

See all 581
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.

  • Vitruvius

    github.com

    An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.