// Agent profile
OpenAI Skills (openai/skills) is the official skills catalog for OpenAI's Codex coding agent. It provides a curated collection of reusable skill packages — folders of instructions, scripts, and resources that AI agents can discover and install to perform specific tasks consistently. Skills are organized into three tiers: .system (auto-installed in latest Codex), .curated (vetted by OpenAI), and .experimental (community-contributed). The repository aligns with the Agent Skills open standard (agentskills.io), establishing a shared format for packaging agent capabilities. With 26K+ GitHub stars, it was a key reference for how major labs approach skill distribution. Note: the repository is now deprecated in favor of the OpenAI Plugins repository, but the patterns and many skills remain influential.
OpenAI Skills is the official catalog of reusable skill packages for the Codex coding agent. Each skill is a self-contained folder of instructions, scripts, and resources that agents can discover, install, and use to complete specific tasks repeatably. The design philosophy — “write once, use everywhere” — treats skills as portable capability packages rather than one-off prompt engineering.
The repository organizes skills into three tiers: .system skills are bundled with the latest Codex versions automatically, .curated skills are vetted and installable by name, and .experimental skills are community-contributed with less formal review. This tiered approach balances ease of discovery with quality control.
Skill installation works through the $skill-installer tool inside Codex. Curated skills install by name ($skill-installer gh-address-comments), while experimental skills can be installed by folder path or direct GitHub URL. Skills take effect after a Codex restart.
Open standard alignment: The repository follows the Agent Skills open standard documented at agentskills.io, establishing a shared format that other agent platforms can adopt for skill packaging and distribution.
The repository is deprecated. OpenAI now directs users to the OpenAI Plugins repository for current Codex extensions, with a separate “Build plugins” guide for custom skill creation. The existing skills and the packaging patterns they established remain a useful reference for anyone designing skill distribution systems for AI agents.
Teams studying how to package agent capabilities for distribution across organizations will find the three-tier model instructive. The curated skills themselves — covering GitHub workflows, code planning, and development patterns — remain functional reference implementations even as the repository moves to maintenance mode.
Codex users who want to extend their agent with pre-built capabilities, and agent platform developers looking for prior art on skill packaging and distribution standards.
Builder.io's open-source framework for building agent-native applications — shared actions, SQL-backed state, identity, tools, skills, jobs, observability, and UI surfaces that all work together.
CLI orchestration tool that installs and configures platform connectors for 16+ social and web sources — Twitter/X, Reddit, YouTube, GitHub, Bilibili, and more. Zero API fees.
Persistent memory layer for AI coding agents — benchmark-backed (95.2% on LongMemEval-S), 92% fewer tokens per session vs full-context pasting, zero manual memory.add() calls.