Skills Officially Comes to Codex

Summary

Agent Skills let you extend Codex with task-specific capabilities. A skill packages instructions, resources, and optional scripts so Codex can perform a specific workflow reliably. You can share skills across teams or the community, and they build on the open Agent Skills standard. Skills are available in both the Codex CLI and IDE extensions. A skill captures a capability expressed through markdown instructions inside a SKILL.md file accompanied by optional scripts, resources, and assets that Codex uses to perform a specific task. my-skill/ SKILL.md Required: instructions + metadata scripts/ Optional: executable code references/ Optional: documentation assets/ Optional: templates, resources Skills use progressive disclosure to manage context efficiently. At startup, Codex loads the name and description of each available skill. Codex can then activate and use a skill in two ways: Explicit invocation: You can include skills directly as part of your prompt. To select one, run the /skills slash command, or start typing $ to mention a skill. (Codex web and iOS don’t support explicit invocation yet, but you can still prompt Codex to use any skill checked into the repo.) Implicit invocation: Codex can decide to use an available skill when the user’s task matches the skill’s description. In either method, Codex reads the full instructions of the invoked skills and any extra references checked into the skill. Codex loads skills from these locations. A skill’s location defines its scope. When Codex loads available skills from these locations, it overwrites skills with the same name from a scope of lower precedence. The list below shows skill scopes and locations in order of precedence (high to low): Skill ScopeLocationSuggested UseREPO$CWD/.codex/skills Current Working Directory: where you launch Codex.If in a repository or code environment, teams can check in skills most relevant to a working folder here. For instance, skills only relevant to a microservice or a code module...