Invocation — handoff routing

Every vt-c-pov verdict ends with a concrete next command. The grade (and, for Adopt, whether scope is clear) picks the route.

Routing table

Verdict	Condition	Handoff
Adopt	scope is clear (you can already name the deliverable)	/vt-c-spec-from-requirements
Adopt	scope is fuzzy (worth doing, shape unclear)	/vt-c-shape
Trial	any	/vt-c-quickfix (time-boxed experiment behind an audit trail)
Hold	any	none — state the re-open trigger (a date, an event, a second incident)
Reject	Tier 3	optional /vt-c-adr — record why we said no so it isn't re-litigated
Reject	Tier 1-2	none required (the verdict text is the record)
Not-our-problem	any	name the target repo/namespace; do not open a toolkit spec

Optional learning capture (any verdict)

Any verdict may additionally hand off to /vt-c-compound to capture the reasoning as a durable learning in docs/solutions/patterns/ (Constitution III). Prefer this when the reasoning generalizes (a reusable adopt/reject heuristic), not just the single decision.

Consuming discovery output

vt-c-pov is designed to sit downstream of the discovery skills:

• /vt-c-repo-evaluate produces "Applicable Patterns" + "Recommended Next Steps" but no graded verdict → it nudges here; vt-c-pov reads its report and grades.
• /vt-c-content-evaluate produces coverage gaps + a weak 3-value signal inside a capture_thought → it nudges here; vt-c-pov grades the adoption question the evaluation raised.

This one-directional flow (discovery → verdict) is the anti-duplication mechanism from Decision 4: the evaluate skills stay discovery-focused, vt-c-pov owns the 5-grade contract. Do not re-implement discovery here, and do not add a competing verdict to the evaluate skills.

Handoff hygiene

• Name the exact command, not a vague "consider planning".
• For Adopt, carry the Project-floor findings into the handoff so the spec/ shape step starts grounded.
• For Not-our-problem, the "handoff" is a pointer to the right repo — never a toolkit spec.