rbrain

Local-first engineering risk intelligence for a single repository.

Before you edit code, rbrain answers:

1. What likely breaks if I change this?
2. Who owns this code today?
3. What depends on this — transitively?
4. What changed recently here?
5. What got riskier this week?
6. What is fragile / weakly tested?
7. What files tend to change together with this one?

Status: V1 in early build. Today only init is implemented; scan, risk, deps, owners, hotspots, audit, ask, create-context arrive in subsequent milestones.

Install

npm install -g rbrain

This installs two equivalent commands: rbrain (short) and repobrain (alias).

Quickstart

cd path/to/your/repo
rbrain init

This creates .repobrain/, writes a default config.json, and appends .repobrain/ to your .gitignore. Idempotent — re-running reports already initialized; no changes.

What it is — and is not

It IS a deterministic engine over your repo state + git history, run locally, with no network calls during a default scan, no repo upload, and no code execution from the analyzed repo.

It is NOT a code generator, NOT code search, NOT a repo chatbot, NOT a linter, and NOT a manager KPI dashboard.

Flags

rbrain [command] [options]

Options:
  --cwd <path>    repository root (defaults to the current working directory)
  --json          emit machine-readable JSON
  --quiet         suppress non-essential output
  --verbose       verbose output
  -V, --version   print the version
  -h, --help      print help

License

MIT.