Every plugin for OpenAI Codex on GitHub — repos with a .codex-plugin/plugin.json or a Codex marketplace. Install with codex plugin marketplace add owner/repo then codex plugin add name@marketplace. Many work in both tools; see Claude Code plugins.
Use MCP servers through the mcporter CLI and saved scripts instead of native MCP tools: a setup agent that moves your servers into mcporter, a session hint, and a hook that points native MCP calls at mcporter.
Make your coding agent fluent in your codebase. A living trace of domain, architecture, and patterns it reads before building — so features fit and nothing gets rebuilt.
SAGE (Session Awareness & Guidance Engine) - a passive, read-only judge for fleets of parallel AI coding sessions. It watches and answers; it never edits, spawns, or blocks.
Build evidence-backed company context and agent-ready GTM ontologies. Three skills: company-context-builder inventories products, ICPs, personas, positioning, and won CRM evidence; gtm-ontology-builder models CRM objects, custom fields, pipelines, automations, agent actions, governance, and KPIs for Pipedrive, HubSpot, Attio, Salesforce, email marketing, and ERP; gtm-context-evaluator runs competency evals against the result. Bundles the renderer, JSON schemas, and a complete worked example.
Blaxel agent skills bundle: SDK skill for creating cloud sandboxes (microVMs) to run code, start dev servers and generate live preview URLs, CLI skill for managing Blaxel resources (agents, sandboxes, jobs, MCP servers) from the command line, and the hosted Blaxel MCP server for managing workspace resources.
Connect your agent to the LlamaParse Platform MCP and unlock access to document processing tools like Parse, Split and Classify, plus agentic retrieval over Index v2 knowledge bases.
Connect your agent to the LlamaParse Platform MCP in the Europe region and unlock access to document processing tools like Parse, Split and Classify, plus agentic retrieval over Index v2 knowledge bases. Use this plugin if your LlamaCloud account lives on cloud.eu.llamaindex.ai; accounts on cloud.llamaindex.ai should install llamaparse-mcp instead.
Drive a dig knowledge base from Claude: the dig skill + the dig mcp server (find/org/drift/reconcile/export/undo). Requires the dig CLI on PATH — https://dig.vllnt.com.
Turn one or more video URLs or local files into a polished Word document with selected screenshots, captions, and timestamped analysis. Can also download caption/subtitle files (SRT, VTT, TXT) on their own.
Community guidelines for empirical SE studies involving LLMs. Use /llm-guidelines:review to check a paper draft and its supplementary material against the eight guidelines, or /llm-guidelines:explore to discuss the guidelines and your study design.
Make SecondBrain your assistant's automatic long-term memory: a skill that recalls what matters and saves durable facts, decisions, preferences, people, and events on its own, plus an auto-memory primer and the SecondBrain connection. Sign-in is a one-time browser login (OAuth) — nothing to paste — in Claude Code, Cowork, claude.ai, and the ChatGPT desktop app / Codex. Backed by your own private notes (Pro).
Migrate PyTorch and Triton kernels to AWS Trainium NKI. Converts nn.Module.forward and @triton.jit kernels (and hand-written CUDA custom kernels) into @nki.jit kernels, then compiles, numerically verifies, and profiles every candidate on a real Trainium device before emitting a diff.
Set up, run, and troubleshoot a self-hosted Dograh voice-AI deployment (Docker, local or remote). Installs prerequisites, drives the official setup scripts, and diagnoses failing services and health checks.