docs: sharpen consent flag behavioral guidance in muse docs
Strengthen --consent documentation from informational to prescriptive:
- unified.md: "requires --consent when called by a proactive agent" on
create, rename, delete
- SKILL.md: add when-to-use context (proactive vs direct-response),
onboarding carve-out on create, and agent double-flag guidance on delete
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>