archdoc

Archdoc Skill

Command: /archdoc

Generate agent-friendly architecture documentation for the currently opened repository.

The skill inspects the repository, fills the architecture documentation templates, removes unused or irrelevant template sections, and writes the final documentation files to the repository’s docs/ folder.

Purpose

/archdoc creates four primary repository documentation files:

docs/REPO_MAP.md
docs/ARCHITECTURE.md
docs/API_SURFACE.md
docs/OPERATIONS.md

These documents help humans and agents understand a codebase quickly and safely.

Responsibility split:

• REPO_MAP.md: repository orientation, important files, commands, conventions, glossary, and agent navigation.
• ARCHITECTURE.md: static architecture, modules, boundaries, dependencies, data ownership, and high-level interface ownership only.
• API_SURFACE.md: detailed public and integration-relevant interfaces such as HTTP endpoints, CLI commands, events, queues, webhooks, exports, schemas, auth rules, and compatibility notes.
• OPERATIONS.md: runtime behavior, local execution, deployment, configuration, debugging, failure modes, health checks, recovery, and operational verification.

The goal is not to produce generic documentation. The goal is to create a practical repository map that helps future agents:

• understand what is in the repo
• find the main entry points
• understand the static architecture
• understand how the system runs
• trace important runtime flows
• identify tests, config, deployment, and failure modes
• avoid inventing patterns that already exist
• make small, safe changes

Slash Command

Use this skill when the user calls:

/archdoc

The skill should generate or update the documentation set directly.

If the user names a specific area, document that area with enough repository context to keep the output accurate. Otherwise, inspect the repository broadly enough to produce a reliable full documentation pass.

Required Output Files