"Build and maintain a layered docs system for durable project knowledge."

This repository contains a specialized skill for AI agents (Claude Code, OpenClaw, Codex) designed to bootstrap, audit, or maintain a repository-specific project docs system centered on AGENTS.md and .docs/.

Durable project knowledge often gets lost in long chat threads. This skill enforces a structured approach to saving reusable knowledge directly in the repository so future sessions can continue with less re-discovery.

Install this skill into your agent environment (Claude Code, OpenClaw, or Codex):

# Using npx skills (Recommended for Claude Code)
npx skills add multicul-silver-wolf/agent-docs-system-skill/skills/project-docs-system

# Using openclaw
openclaw install https://github.com/multicul-silver-wolf/agent-docs-system-skill/skills/project-docs-system

# Or clone and install locally
git clone https://github.com/multicul-silver-wolf/agent-docs-system-skill
openclaw install ./agent-docs-system-skill/skills/project-docs-system

The docs system uses a two-level domain structure with index maps at each folder level:

• AGENTS.md: Agent entry point. Adds docs-system rules only after .docs/ is initialized.
• .docs/DOCS.md: Cross-domain durable conventions and recurring preferences.
• .docs/index.md: Top-level map of first-level domains.
• .docs/<domain>/index.md: Map of second-level docs inside one domain.
• .docs/<domain>/<subdomain>.md: Durable docs for one subdomain.

Trigger the skill in your chat:

• "Bootstrap the project docs system in this repo"
• "Update .docs/DOCS.md with our global API naming conventions"
• "Audit .docs/ and fix missing index links"
• "Add a new subdomain doc under .docs/frontend/ and update maps"

Built by Silver Wolf and Sawana.
Inspired by the Aether Editing philosophy.