Ronan Skills (ro)
A Claude Code plugin bundling 29 personal skills for development, quality, browser/visual, audio/media, research, and project setup. Invoke any skill as /ro:<skill-name>.
Also publishable to Cursor, and individual skills work with 40+ other AI agents via npx skills add.
Install
Claude Code (recommended — bundles all 29 skills in one command)
/plugin marketplace add RonanCodes/ronan-skills
/plugin install ro@ronan-skills
Skills then appear as /ro:ralph, /ro:commit, /ro:tdd, etc. Run /plugin to manage.
Cursor
The same repo is also a Cursor plugin (see .cursor-plugin/plugin.json). Submit via cursor.com/marketplace/publish or point Cursor at the skills/ directory locally. Cursor CLI does not yet support plugins — IDE only.
Other agents (Codex, Cline, etc.) — per-skill install
npx skills add RonanCodes/ronan-skills/skills/ralph -g
npx skills add RonanCodes/ronan-skills/skills/commit -g
# ...etc per skill, -g for global, omit for project-local
Configuration
Skills that need API keys read from a shared env file via the ro meta-CLI (lives in skills/context/scripts/ro, symlinked into ~/.local/bin/ro so it's invokable as ro from any terminal). See /ro:context for the full reference.
Single context (most users)
~/.claude/.env.personal # all your credentials (mode 600)
~/.claude/.env # symlink to .env.personal
Skills source via $(ro context env) which resolves to ~/.claude/.env.personal automatically. See .env.example for every variable, where to generate it, and which skill consumes it. First-time setup:
ro context init # creates the template + symlink
# then edit ~/.claude/.env.personal and fill in your keys
# or: /ro:setup-wizard --tokens for a guided walkthrough
Multiple contexts (work / personal / per-client)
If you have separate credentials per project (e.g. a personal Cloudflare account + a client's Cloudflare account), ro switches between them with a three-tier resolver:
$RO_CONTEXT env var — one-shot override per shell~/.claude/contexts.json:active — manual override set by ro context use <name>.ro-context.local in cwd or any ancestor — gitignored contributor override.ro-context in cwd or any ancestor — committed, repo's declared context (cleanest)- cwd-glob rule in
~/.claude/contexts.json — e.g. ~/Github-Acme/** → acme
default in ~/.claude/contexts.json
# Add another context
cp ~/.claude/.env.personal ~/.claude/.env.acme
# …edit with the client's credentials…
# Pin a repo to its context (preferred — committed, no global config needed)
echo acme > /path/to/repo/.ro-context
# Or auto-resolve via cwd-globs
ro context add-rule '~/Github-Acme/**' acme
# See what resolves and why
ro context where # context = acme (resolved from: .ro-context (...))
ro context env # /Users/you/.claude/.env.acme
# Manual override (sticks until cleared)
ro context use acme
ro context auto # back to cwd-resolution
Backwards compat: skills that haven't migrated to $(ro context env) keep working via the ~/.claude/.env symlink, which ro context use updates.
Skills
All skills invoke as /ro:<skill-name> in Claude Code.
Development Workflow
| Skill | Description |
|---|
| ralph | Autonomous build loop. Picks tasks from .ralph/prd.json, implements, validates, commits. |
| write-a-prd | Generate a PRD through an interactive interview. Quick or plan mode. |
| tdd | Test-driven development with red-green-refactor cycles and vertical slices. |
| commit | Emoji conventional commit format. Handles staging, messages, timestamp rules. |
| close-the-loop | Verification loop — tests pass, UI works, screenshots match. |
| debug-escape | Break out of debugging loops by stepping back and researching. |
| post-mortem | Document a resolved bug as a structured post-mortem. |
| coding-principles | KISS, SOLID, DRY, tracer bullets. Index always loaded, detail files on demand. |
Quality & Review
| Skill | Description |
|---|
| grill-me | Stress-test plans, designs, PRDs, or code with relentless probing questions. |
| ubiquitous-language | DDD-style glossary for consistent domain terminology. |
| git-guardrails | Blocks destructive git commands, suggests safer alternatives. (auto-loaded) |
| security-audit | Pre-publish safety check — secrets, PII, work info, risky git history. |
Browser & Visual
