| name | rb:intro |
| description | Walking a user through the Ruby/Rails/Grape plugin commands, capabilities, and workflow. Tutorial-style intro for newcomers who want to learn what the plugin offers rather than tackle a specific task. |
| argument-hint | [--section N] |
| effort | low |
| disable-model-invocation | true |
Intro
Interactive introduction to the Ruby/Rails/Grape plugin for new users.
Quick Start
- Plan a feature:
/rb:plan Add user authentication with Devise
- Work the plan:
/rb:work .claude/plans/user-authentication/plan.md
- Verify & Review:
/rb:verify then /rb:review
What You Get
- 19 specialist agents: ActiveRecord, Hotwire, security, Sidekiq, provenance experts
- 52 skills: Commands for every phase of development
- 22 Iron Laws: Non-negotiable rules enforced automatically
Installation Notes
Marketplace Install (Important)
When installing from the Claude Code marketplace, the plugin's specialist agents will prompt for permissions on first use.
This is a security restriction — plugin agents follow your session permission policy, so project-level permissions.allow rules are the normal fix.
Workarounds:
- Add permissions to your project's
.claude/settings.json:
- Command rules:
Bash(bundle *), Bash(rails *), Bash(rake *),
Bash(mkdir -p **/.claude/**),
Bash(*/bin/manifest-update *),
Read(*), Grep(*)
- Recursive Edit rules for plugin artifact namespaces (
Edit(path)
covers the Write tool; scoped Write(path) rules emit a startup
warning and are not consulted by the file-permission check):
Edit(**/.claude/plans/**), Edit(**/.claude/reviews/**),
Edit(**/.claude/audit/**), Edit(**/.claude/research/**),
Edit(**/.claude/solutions/**),
Edit(**/.claude/skill-metrics/**),
Edit(**/.claude/investigations/**)
- Run
/update-config to apply the recommended Edit allowlist without hand-editing
- Run
/rb:permissions to generate a narrower project allowlist from recent usage
- Use
--plugin-dir for local development while iterating on the plugin itself
- Set
CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1 in your shell. Enables
SendMessage so spawn-fanout skills can resume agents that paused
at maxTurns. Without it, paused agents become coverage gaps.
See CLAUDE.md "Conventions → Agents" section for details.
Detailed Tutorial
See ${CLAUDE_SKILL_DIR}/references/tutorial-content.md for the complete walkthrough with:
- Core workflow commands and decision guide
- Knowledge & safety net (Iron Laws, auto-loading)
- Hooks & behavioral rules explanation
- Command cheat sheet and best practices
- Claude Code built-in features (
xhigh, /focus, /recap,
/less-permission-prompts, /output-styles)
- Keeping
CLAUDE.md small + scoped-rule pattern
Related — invoke manually if needed
- Feature still fuzzy →
/rb:brainstorm (workflow on-ramp; pre-plan discovery)
- Pattern reference needed →
/rb:examples (codebase pattern surface)
