Quick Start

1. Initialize

Create a .stato/ directory in your project:

stato init

This creates:

.stato/ directory for your modules
.stato/prompts/crystallize.md (coding agent prompt)
.stato/prompts/crystallize_web.md (web AI prompt)
.statoignore template for suppressing privacy scan false positives

2. Crystallize

Save a prompt that tells your coding agent how to capture its expertise:

stato crystallize

This saves the prompt to .stato/prompts/crystallize.md. Tell your coding agent: “Read and follow .stato/prompts/crystallize.md”. The agent reads the file and writes stato modules to .stato/.

For web AI (Claude.ai, Gemini, ChatGPT), use:

stato crystallize --web | pbcopy

See Crystallize for details on the two modes.

3. Validate

Check that the agent wrote valid modules:

stato validate .stato/

The 7-pass compiler catches syntax errors, missing fields, type mismatches, and semantic issues. Auto-corrections fix common problems silently.

4. Generate a Bridge

Create a bridge file that points your coding agent to the expertise modules:

stato bridge --platform claude

This generates CLAUDE.md, which Claude Code reads automatically. The bridge is a lightweight table of contents (~500 tokens) that tells the agent what skills are available and where to find them.

Other platforms:

stato bridge --platform cursor    # .cursorrules
stato bridge --platform codex     # AGENTS.md
stato bridge --platform generic   # README.stato.md
stato bridge --platform all       # all four

5. Check Status

See all modules, plan progress, and warnings:

stato status

Output shows a table of all modules with their type, class name, and version, plus plan progress panels for any plan modules.

Full Workflow Example

# Initialize
stato init

# Save crystallize prompt, tell agent to read it
stato crystallize

# Validate
stato validate .stato/

# Generate bridge
stato bridge --platform claude

# Check what you have
stato status

# Export for sharing
stato snapshot --name "my-expertise" --template