How to use
- Copy the installation command and review the source package before installing it in your own agent environment.
- Check the source document for required tools, credentials and supported hosts.
- Describe your input and desired deliverable; review the output against the source material.
npx skills add https://github.com/yeachan-heo/oh-my-claudecode --skill mapBefore you start
- Installation does not include a model subscription, API allowance or permission to access your files.
- Repository stars and installation counts describe activity, not an independent quality or security test.
Rendered from the source repository. Headings, examples, code, tables, links, and referenced images are preserved.
Map
You know something needs doing; you do not know which skill owns it. Find the job below and follow its route. Most routes call the Skill tool; a route marked manual gives you the slash command to invoke yourself because model invocation is disabled. The map routes and explains nothing twice — each routed skill owns its own doctrine.
The delivery loop
| You are here | Skill |
|---|---|
| External work arrived (issues, PRs, reports) | harbor |
| A non-engineer handed you a chat log as requirements | intent |
| The destination is foggy — the first three decisions cannot be stated yet | ask-navigator |
| Vague idea; questions must come before anything else | deep-interview |
| Requirements are clear; you need a plan | plan (multi-perspective consensus: ralplan) |
| Approved plan; decompose and ship end to end with human checkpoints | launch |
| One task; persistence until verified complete | ralph |
| Hands-off idea-to-code | autopilot |
| Work needs coordinated parallel workers | team |
| One approved task carried through verified code | execute |
| One attempt risks locking in the wrong shape | arena |
| Code crosses a function boundary; designs should compete first | architect |
| Finished work needs attack before it ships | interrogate |
| A design question prose cannot settle | loft |
| Finished work needs judgment | review |
| Something claims to be done | verify |
| Work is landing as a PR | pr |
| An open PR needs driving to merge-ready | babysit |
| The run taught the environment something | refit |
On-ramps (off the main loop)
| Job | Skill |
|---|---|
| A bug needs its root cause | debug; competing-hypothesis tracing: trace |
| The question is why the code is this way | why |
| A subsystem needs explaining | how; as a lesson: teach |
| A risky change needs its blast radius mapped | blast-radius |
| The same agent mistake keeps recurring | correct |
| Long autonomous work needs an audit trail | show-me-your-work |
| No bundled skill fits the task shape | figure-it-out (designs one, validated against anchors) |
| A property tests cannot reach needs formal proof | create-formal-verification-skill; drifted spec: maintain-formal-verification-skill |
| A flow or protocol bug shows up as a path through states | statechart |
| The repo needs its own verification commands recorded | create-verification-skill; drifted commands: maintain-verification-skill |
| Roles should run on chosen models | models |
| Tests should drive the implementation | tdd |
| The codebase is getting harder to change | architecture-survey |
| AI-generated slop needs cleanup | ai-slop-cleaner |
| A measured number needs vetting before it is reported or acted on | benchmark-checklist |
| A delegated brief needs cross-cutting judgment guardrails | principles |
| An open question needs grounded answers | research; external docs: external-context |
| Knowledge should compound across sessions | wiki; durable project memory: remember |
| A recurring autonomous improvement loop | self-improve; durable goal ledger: ultragoal |
| A session needs reorientation across its pivots and outcomes | session-flow — run /oh-my-claudecode:session-flow manually |
| Visual work needs verdicts or diagrams | visual-verdict; diagram |
Utilities
| Job | Skill |
|---|---|
| Install, diagnose, configure | omc-setup; omc-doctor; configure-notifications |
| Manage skills themselves | skill (local); skillify (extract from session) |
| Inspect orchestration state | trace (also the tracing lane); graph (DAG runtime) |
| Cancel a mode | cancel |
| CI is red on a PR you own | fix-ci; review threads: get-pr-comments; conflicts: fix-merge-conflicts |
| The user missed the session | ketchup; as a subtitled video: ketchup-video; after a break or crash: recall |
| A brief is vague | sharpen; playback before starting: in-your-own-words |
| A reply was too dense | bro |
| Shipping prose needs the standard | technical-writing; TS type discipline: typescript-best-practices |
| The period's authored work needs summarizing | what-did-i-get-done |
| A queue must land as one stacked PR chain | stacked-delivery; watch state: babysit |
| A measured target needs climbing over multiple rounds | hillclimb |
| A long session's lessons should be banked as skill edits | reflect; extraction: skillify |
| A cross-project personal behavior profile is wanted | personal-mode; single facts: remember |
Rules
- The map routes; it never executes. Each ordinary row resolves to one Skill-tool call — make it, then follow that skill's doctrine. For a manual-only route, show the user its slash command; do not try to call the Skill tool.
- When two rows both fit, the delivery loop above is the tiebreak: earlier legs come first.
- Phase boundaries (continue / re-enter from disk / hand off / compact) are defined in
launch's Context hygiene — the map does not restate them. - A job matching no row is itself information: say so plainly rather than picking the nearest-sounding skill.
