Skip to main content

​
init

Bootstraps a repo end to end: system-doctor preflight, repo analysis (code repos), git hooks, and agent context. Idempotent — re-running reports “Already initialised” and writes nothing. 
tiny-brain init
tiny-brain init --type prose --yes     # writing repo, skip analysis + prompt
tiny-brain init --path ./my-repo       # initialise another directory
tiny-brain init --ignore-system-doctor # push past preflight warnings (never criticals)

​
analyse

Detects the repo’s languages, frameworks, package manager, scripts, and test patterns, writing the result to .tiny-brain/analysis.json (and technology context files under .tiny-brain/tech/). 
tiny-brain analyse
tiny-brain analyse /path/to/repo
Agents read analysis.json to run your scripts (npm run test, not an improvised test runner) — re-run analyse after changing the toolchain.

​
doctor

Toolchain health checks. Exits 0 when everything passes or warns, 1 on any critical. Each failing check prints a copy-pasteable Fix: line. 
tiny-brain doctor
tiny-brain doctor --migrate   # run registered source-of-truth migrations (idempotent)

​
config

Credentials and preferences. Preferences resolve repo → global → default; show-sources shows which layer each value came from. 
tiny-brain config preferences list
tiny-brain config preferences get enableTDD
tiny-brain config preferences set defaultPersona developer --global
tiny-brain config preferences show-sources
tiny-brain config set clientId YOUR_CLIENT_ID     # account credentials
tiny-brain config show                            # credentials, masked
Repo preferences live in .tiny-brain/config.json; global ones in ~/.tiny-brain/config/preferences.json. Notable keys: enableTDD, enableQuality, defaultPersona, testPatterns, reviewPipeline, and the directory keys (docsDirectory, prdDirectory, adrDirectory, fixesDirectory). An unknown key errors with the full valid-key list, so get is a safe way to probe.