Bootstrap a repo

The bootstrap plugin reads your code and docs, identifies domains and languages, and generates agent prompts that know your code. Run it in Claude Code in your terminal, not in the Desktop app or the web app.

Prerequisites

• the Closedloop code plugin installed (Install the plugins), plus bootstrap: claude plugin install bootstrap@closedloop-ai --scope user
• a project with at least a README.md; CLAUDE.md and ARCHITECTURE.md help a lot

Run it

/bootstrap:agent-bootstrap --depth medium --strategy backup

Useful flags

• --depth quick|medium|deep — how deeply to explore (default medium).
• --focus <area> — constrain analysis to one area (frontend, backend, infra, mobile, web).
• --dry-run — preview what would be generated without writing files.
• --update — regenerate outdated agents based on project changes.
• --strategy backup|skip|overwrite|interactive — conflict strategy (default backup).
• --add-domain <domain> — add agents for a specific domain.

What you get

• A set of .claude/agents/*.md tailored to your codebase.
• .closedloop-ai/settings/critic-gates.json that the code plugin uses to pick critics.
• .closedloop-ai/bootstrap-metadata.json, which --update uses to track generated agents.
• A run directory at .closedloop-ai/bootstrap/<timestamp>/ with bootstrap-report.md and validation-report.json.

When to re-run

• When the project's architecture changes materially.
• When you add a new domain (--add-domain).
• When you want to refresh agents against new conventions (--update).
• On demand with --dry-run to compare what the generator would produce to the current agents.

After bootstrap, run your first loop with /code:code in Claude Code to validate the agent suite against a real task.