Skills
Introduces all 37 OMC skills organized into workflow and utility categories.
What are Skills?
Skills are commands invoked with the format /oh-my-claudecode:skill-name.
They can also be activated via magic keywords, so you can use them in conversation without memorizing the command name.
Count note: OMC ships 37 built-in skills. 36 are user-callable via slash command or magic keyword; 1 (
omc-reference) is auto-loaded internally and hasuser-invocable: false.
Workflow Skills
Execution mode skills that automate tasks.
| Skill | Description | Magic Keywords |
|---|---|---|
| autopilot | Autonomous full pipeline from idea to working code | autopilot, build me, I want a |
| ralph | Self-referential loop that repeats until completion | ralph, don't stop |
| ultrawork | Maximum parallel agent execution | ultrawork, ulw |
| ultraqa | QA cycling: test, verify, fix, repeat | - |
| team | N Claude agent coordination pipeline | - |
| omc-teams | Run claude/codex/gemini CLI workers in tmux | - |
| ccg | Claude-Codex-Gemini tri-model orchestration | ccg |
| plan | Interview-based strategic planning | - |
| ralplan | Planner-Architect-Critic consensus planning | ralplan |
| sciomc | Parallel scientist agent research orchestration | - |
| external-context | Parallel document-specialist external documentation lookup | - |
| deepinit | Hierarchical AGENTS.md codebase documentation | - |
| deep-interview | Socratic deep interview with mathematical ambiguity gating | deep interview, ouroboros |
| deep-dive | 2-stage pipeline: trace investigation → deep-interview requirements crystallization | deep dive |
| self-improve | Autonomous evolutionary code improvement engine with tournament selection | - |
Utility Skills
Configuration, management, and helper skills.
| Skill | Description | Magic Keywords |
|---|---|---|
| cancel | Cancel all active modes | cancelomc, stopomc |
| ask | Ask Claude/Codex/Gemini and capture artifacts | - |
| ai-slop-cleaner | Clean up AI code slop | deslop, anti-slop |
| learner | Extract reusable skills from the current session | - |
| skill | Manage local skills (list, add, remove, search) | - |
| trace | Evidence-driven tracing lane | - |
| hud | HUD display configuration | - |
| configure-notifications | Notification integrations (Telegram, Discord, Slack) | configure telegram, configure discord, configure slack |
| release | Automated release workflow | - |
| writer-memory | Agentic memory system for writers | - |
| project-session-manager | git worktree + tmux isolated development environment | - |
| visual-verdict | Structured visual QA verdict for screenshot-to-reference comparisons | - |
| omc-setup | OMC setup and configuration | - |
| mcp-setup | MCP server setup | - |
| omc-doctor | Diagnose and fix installation issues | - |
| setup | Unified setup entry point | - |
| debug | Diagnose OMC session, hooks, state, and orchestration | - |
| verify | Evidence-based verification before claiming completion | - |
| remember | Route knowledge to project memory, notepad, or docs | - |
| skillify | Turn a session workflow into a reusable SKILL.md draft | - |
| wiki | Persistent markdown knowledge base (Karpathy LLM Wiki) | wiki, wiki add, wiki query |
| omc-reference | Auto-loaded OMC catalog reference (internal) | - |
How to Invoke Skills
Slash Commands
/oh-my-claudecode:autopilot "REST API for a bookstore"
/oh-my-claudecode:ralph "fix all TypeScript errors"
/oh-my-claudecode:team 3:executor "implement user auth"Magic Keywords
Keywords typed in conversation activate the skill automatically.
autopilot build me a todo app with React
ralph fix the authentication module
ulw implement user managementSkill Relationship Map
deep-interview (clarify requirements)
→ ralplan (consensus planning)
→ autopilot (autonomous execution)
→ ralph (loop execution)
→ ultrawork (parallel execution)
→ ultraqa (QA cycling)Higher-level skills automatically include lower-level skills. Running autopilot internally enlists ralph and ultrawork to help.