OpenAI Plugins
github.com/openai/plugins- Category
- Developer Tools
- Rank
- No. 808Tools index
- Type
- APP
- Builder
- openai
- GitHub
- 6.8k stars
- Date
About
OpenAI's repository of official plugins.
What it does
A collection of Codex plugin examples that bundle skills, connectors, MCP servers, commands, agents, hooks, and assets. Individual plugins can expose only the surfaces they need, from simple connector routing to local evaluation and security workflows.
Why it's ranked here
The collection is useful because it shows several concrete plugin shapes instead of one narrow template. Examples cover design tools, productivity services, mobile and web development, deployment, video, presentations, evaluation, and security. Its value is breadth and working implementation detail, not a single unified product experience.
What's good
The examples demonstrate meaningful implementation patterns. The evaluation tool supports analysis, budget explanation, measurement planning, benchmarks, comparisons, and three report formats. The developer MCP server validates environment names, confines destinations to a workspace, and requires explicit approval before directing a key write.
Tradeoffs
Consistency varies because the repository is a collection of separate bundles. Some examples are minimal connector manifests, while others contain substantial command-line or MCP implementations. The supplied text documents an MIT license only for the Expo plugin, so it does not establish one repository-wide license.
How to use it well
Use it when building or studying Codex plugins, especially when you need examples of skills, connectors, MCP servers, local evaluation, or approval-gated workflows. Start from the example closest to your integration shape. It does not replace generic web search, broad analytics ownership, or standalone product documentation.
Technical notes+
README.md describes a plugin-per-directory structure with required manifests and optional skills, connectors, MCP configuration, agents, commands, hooks, and assets. plugins/plugin-eval/src/cli.js implements local analysis, workflow guidance, budget explanation, measurement plans, benchmark initialization and execution, reporting, and comparisons. plugins/plugin-eval/src/index.js exposes the evaluation capabilities as library exports, while plugins/plugin-eval/src/renderers/index.js renders JSON, Markdown, or HTML. plugins/openai-developers/mcp/server.mjs implements a line-oriented JSON-RPC MCP server with workspace path confinement, environment-name validation, and elicited destination approval. plugins/codex-security/mcp/server.mjs reconstructs and executes a Brotli-compressed runtime. plugins/superhuman/bin/cli.js launches a remote MCP endpoint through npx.
Observed
- Repository structure
- Curated Codex plugin examples organized as separate plugin bundles.
- Implementation language
- The supplied CLI, library, and MCP implementation files use JavaScript.
- Interfaces
- Examples include command-line tools, library exports, MCP servers, connector manifests, and skills.
- Evaluation output
- The plugin evaluation tooling renders JSON, Markdown, and HTML.
- Runtime surface
- Supplied executable implementations use Node.js modules and APIs.
- License evidence
- The Expo plugin includes the MIT License; no repository-wide license is established by the supplied text.
Read from README.md, plugins/superhuman/bin/cli.js, plugins/plugin-eval/src/cli.js, plugins/plugin-eval/src/index.js, plugins/codex-security/mcp/server.mjs, plugins/openai-developers/mcp/server.mjs, plugins/plugin-eval/src/renderers/index.js, plugins/expo/LICENSE, plugins/box/.app.json, plugins/fal/.app.json, plugins/hex/.app.json, plugins/hex/README.md, plugins/jam/.app.json, plugins/brex/.app.json, plugins/clay/.app.json.
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.