| name | ui-flow-offender-audit |
| description | Detect prompt-chain offenders in ytnova UI flows and produce ranked compression plans that enforce the UX economy gate. |
UI Flow Offender Audit
Use this skill when auditing interaction flow depth, prompt chaining, or menu-dive friction.
Audit Objective
Find flows where the common path violates UX economy:
- Common-path target:
key -> Enter -> result
- Maximum common-path submenu depth: 1
Workflow
- Locate interaction chains in code/tests (menus, prompts, modal choices).
- For each flow, capture:
- current chain
- common-path submenu depth
- common-path decisions before result
- Flag offenders where depth > 1 or where defaults cannot reach fast path.
- Propose compressed flow preserving behavior and safety.
- Rank offenders by user-frequency impact and chain depth.
Required Output
- Ranked offender table with:
- command/flow name
- current chain
- proposed chain
- current vs proposed submenu depth
- expected speed gain
- likely files/tests to update
- Per-offender implementation prompt suitable for a clean-slate AI.
Guardrails
- Do not remove safety confirmations for destructive operations without replacement safeguards.
- You MUST combine decisions into one prompt instead of adding nested submenus.
- Keep advanced options available as in-prompt toggles/defaults.