Skills Overview

Skills should package repeatable Codex workflows so that repository instructions and docs do not become overloaded with procedural detail.

flowchart LR
  request["Recurring task"] --> skill["Relevant SKILL.md"]
  skill --> references["References"]
  skill --> scripts["Optional scripts"]
  skill --> codex["Focused Codex execution"]
  codex --> artifact["Plan, code, docs, or verification output"]

Skill Categories

Category	Purpose	Example Future Skills
Engineering	How to safely perform recurring code or docs work.	large-change, mkdocs-update, testing
Product	How to reason about Anuva product concepts.	architecture-review, capability-mapping
Repository	How to work inside a specific repository type.	repository-bootstrap, docs-publishing

First-Draft Policy

Do not create the full skill catalog yet. This handbook should first define the architecture, registries, and templates that future skills will use.

Relationship To Docs And Agents

• Docs explain what is true.
• AGENTS.md explains how Codex should behave in a repository.
• Skills explain how Codex should perform a recurring task.