# Vibe What Do I Want

> Clarify intent and freeze requirements by entering the canonical vibe runtime with a discovery-first bounded stop.

- Skill: `zuoguyoupan2023/vibe-what-do-i-want` (Agent Skill)
- Install (CLI): `npx skillmds@latest add zuoguyoupan2023/vibe-what-do-i-want`
- Raw SKILL.md: https://api.skillmd.com/api/skills/zuoguyoupan2023/vibe-what-do-i-want/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Product & Planning
- Author: zuoguyoupan2023 (https://skillmd.com/u/zuoguyoupan2023)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/zuoguyoupan2023/vibe-what-do-i-want

---


Use canonical `vibe` as the only governed runtime authority for this request.

Entry bias:

- clarify what the user actually wants
- shape scope before solutioning
- strengthen requirement-first discovery
- stop after canonical `requirement_doc`

Execution rules:

- delegate to canonical `vibe`
- keep `vibe` as the only runtime authority
- let canonical `vibe` keep router selection and `confirm_required` ownership
- do not create a second requirement surface
- do not create a second plan surface
- do not create a parallel runtime
- do not preflight-scan the current workspace or repository for canonical proof files before launch
- Launch canonical-entry first; validate receipts only after it returns a session root
- do not continue into `xl_plan`, `plan_execute`, or `phase_cleanup` unless the user explicitly re-enters through canonical `vibe` or another approved wrapper
- when a later wrapper intentionally continues from this bounded stop, forward `--continue-from-run-id <source_run_id>` and `--bounded-reentry-token <reentry_token>` from the latest `runtime-summary.json` `bounded_return_control` block

When this wrapper is chosen, enter canonical `vibe` with:

- stronger `deep_interview`
- stronger requirement clarification before planning
- a bounded terminal stage of `requirement_doc`

Request:
$ARGUMENTS