AIAny
AI Agent2026
Icon for item

OpenAI Plugins

A curated collection of Codex plugin examples demonstrating plugin manifests, companion surfaces (skills, hooks, assets), and sample integrations. Highlights richer, opinionated examples for Figma, Notion, iOS/macOS/web builds and MCP-backed bundles — useful for prototyping assistant plugins.

Introduction

Most teams building assistant features stall not on models but on the integration surface: manifests, action schemas, and companion UIs. This repository bundles concrete Codex plugin examples so teams can skip specification debates and iterate on real plugin surfaces that connect LLMs to tools, UIs, and deployment workflows. (Repository metadata: created 2026-03-04; ~1k stars.)

What Sets It Apart
  • Concrete manifest-first layout: each plugin lives under plugins/<name>/ with a required .codex-plugin/plugin.json plus optional skills, commands, hooks, and assets — so you get a runnable example of the full integration surface rather than an abstract spec.
  • Rich, opinionated end-to-end examples: bundles for Figma, Notion, iOS/macOS/web apps, Expo, and MCP-backed workflows — so you can see UI, backend, and skill wiring together instead of isolated endpoints.
  • Focus on developer ergonomics: companion surfaces (skills/, .app.json, agents/, commands/) show typical patterns for authentication, input/output shaping, and user-facing prompts — so adapting the pattern to your product is faster.
  • Curated, example-first approach: emphasizes copy-pasteable patterns and trade-off notes rather than exhaustive platform APIs — so it’s efficient for prototyping and learning common pitfalls.
Who It's For & Trade-offs

Great fit if you are a developer or team prototyping assistant plugins and want runnable, end-to-end examples that show manifest shape, UI wiring, and backend hooks. It accelerates feature validation and cross-team alignment (product ↔ engineering ↔ design).

Look elsewhere if you need production-ready, hardened plugin middleware, compliance/audit tooling, or a turnkey plugin marketplace — this repo provides examples and patterns, not a managed deployment platform or formal security review. It also targets the Codex/plugin manifest workflow specifically, so teams using a different plugin spec may need translation work.

Where It Fits

Think of this repo as the cookbook of integration patterns between LLMs and external tools: not a runtime or platform, but a repeatable set of blueprints for common integrations (design tools, note apps, build workflows, mobile/web packaging).

How It’s Organized (brief)

Plugins are organized under plugins/<name>/ with a manifest (.codex-plugin/plugin.json) and optional companion folders (skills/, assets/, agents/, commands/, .app.json). That structure makes it straightforward to extract a manifest and the minimal companion code you need to prototype an assistant-enabled feature.

Overall, the repo is most valuable as a learning and prototyping resource: copy the example that matches your use case, adapt authentication and I/O, and iterate on UX and authorization before investing in production infrastructure.

Information

  • Websitegithub.com
  • AuthorsOpenAI
  • Published date2026/03/04

Categories

More Items

Hugging Face
AI Model2026

Open-weights LLM fine-tuned for phone-based voice agents that prioritizes low latency and reliable tool/function calling. Based on NVIDIA Nemotron 3 Nano (30B total, 3.5B active), supports very long contexts (262,144 tokens) and recommends temperature=0 with thinking disabled for deployment.

GitHub
AI Agent2026

A curated collection of production-ready Agent Skills that turn tasks—presentation production, image generation, local KB retrieval, article assembly, and web-design—into agent-loadable skill folders. Uses a SKILL.md contract, supports multiple agent runtimes (Claude Code, Cursor, Codex), and offers modular install paths with pinned release zips.

GitHub
AI Agent2026

Provides a ruleset and skills that make AI coding agents prefer the simplest correct implementation: reuse existing code, prefer stdlib/native features, and only write minimal new code. Cuts generated LOC, tokens, cost and time while preserving validation and safety.