audit
Audit current renders and stress coverage, repair real defects and stop under the bound refinement policy.
Dispatch the existing candidate-generation stage from minimal tokens to one authentic direction choice.
$ npx -y skills add 3x-haust/oh-my-design --skill omd-concepts --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/omd-conceptsContext preview
The summary Claude sees to decide when to auto-load this skill.
Dispatch the existing candidate-generation stage from minimal tokens to one authentic direction choice.
name: omd-concepts description: Dispatch the existing candidate-generation stage from minimal tokens to one authentic direction choice.
Read current stage work and the concept procedure only. Art Director commits a minimal seed and candidate-plan-v2 before independent Sketch source. Default two concepts, three for a named uncertainty: one representative surface, same real content/state/views, different structural/density/type relationships. The rest of Frame scope stays required.
Use schema candidate-content/plan/set and candidate plan/check/packet. Capture actual previews through the native measurement path; final production floors do not gate provisional studies. Present every preview ID and tradeoff. Interactive runs wait for ONE real user choice through candidate select; bare continue selects nothing. Explicit autonomy requires fresh isolated Eye and a real grant, never a model-authored flag.
Run candidate check --current. Stale consumed inputs return to the same checkpoint. Missing viable evidence remains honest debt, not approval or a universal first-render blocker. Then dispatch selected type/system work; no second tournament.
AI that designs like a human. Give OMD one request. It researches, compares real material, chooses a direction, builds, renders, and repairs internally.
Audit current renders and stress coverage, repair real defects and stop under the bound refinement policy.
Interpret recurring measured findings and honest trends without turning taste into a skill score.
Dispatch the existing candidate-generation stage from minimal tokens to one authentic direction choice.
Review actual current design evidence and rank observable problems without implementing changes.
Expand the selected sample into the complete requested surface, state and viewport scope.