bw-analyse
01Command Claude Code
Analyse an unfamiliar or brownfield codebase and write structured context docs. Use when codebase docs or project technology steering are missing or stale.
Approve intent, ship autonomously. Agent-first autonomous development workflow for Claude Code, OpenCode, and OpenClaw — you read a one-page spec and say yes, the agent handles research, planning, build, security scan, code review, and PR. Everything else is automated. Works on any project via a single curl | bash
This repository also configures its own agents. See what buildwright tells them →
Command Claude Code
Analyse an unfamiliar or brownfield codebase and write structured context docs. Use when codebase docs or project technology steering are missing or stale.
Command Claude Code
Research a question or topic and produce a written deliverable without implementation or commits. Use when analysis or an implementation-ready plan is needed.
Command Claude Code
Independently review a PR or local changes for code and security issues without modifying code. Use before shipping or when a standalone review is requested.
Command Claude Code
Run the full quality pipeline, then commit, push, and create a PR. Use when verified work is ready to be reviewed and shipped.
Command Claude Code
Run the project's typecheck, lint, test, and build gates. Use for quick validation before review or shipping.
Command Claude Code
Implement bug fixes, refactors, and features with research, TDD, documentation, verification, and review. Use when a request needs code or project files to change (build, add, fix, change, or refactor), even when the user does not type /bw-work.
Command Claude Code
Sweep a repository for stale docs, dangling references, dead files, duplicated rules, and ineffective gates. Use for periodic repository hygiene outside a feature diff.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: