| name | spk-run-verdict-capture |
| description | Set a Spec Kitty work-package review verdict through the deterministic event-log seam, so every agent harness records approve/reject the same way. |
spk-run-verdict-capture
Use this skill whenever an agent (any of the 13 slash-command harnesses or the shared
Agent-Skill harnesses) needs to record the outcome of a WP review — approve, reject, or
otherwise. It documents the single deterministic seam so verdicts never diverge by harness.
The authoritative seam
The sole authority for a WP's review verdict is the review_result event in the mission's
status.events.jsonl (read back via specify_cli.status.event_sourced_review_result). The
review-cycle-N.md render is non-authoritative prose — do not hand-edit it to change a
verdict, and never treat it as the source of truth. Because the verdict is an event, it is
append-only, deterministic, and identical across every agent.
How to set a verdict (use the CLI — do not write the event by hand)
- Approve a WP under review:
spec-kitty agent tasks move-task <WP> --to approved --mission <slug> --note "<why>"
- Reject a WP (send it back with structured feedback):
spec-kitty agent tasks move-task <WP> --to planned \
--review-feedback-file <path/to/review-feedback-N.md> --mission <slug>
- Terminal (done) with an explicit review triple (reviewer / verdict / reference):
spec-kitty agent status emit <WP> --to done --actor <name> \
--evidence-json '{"review": {"reviewer": "<name>", "verdict": "approved", "reference": "<PR/commit>"}}'
Each command emits the review_result (or review-bearing transition) event; the reduced snapshot
is what spec-kitty agent tasks status and merge preflight read.
Verdict vocabulary
The verdict field is validated against the canonical bridge specify_cli.status.event_verdicts()
plus the proof-event extras commented, rejected, unknown (see
specify_cli.proof.events). Do not invent verdict strings; an unknown verdict is rejected at the
seam. Moving a WP to done requires a review triple with reviewer, verdict, and reference.
Guardrails
- One seam, every harness: Codex, Claude, Gemini, Copilot, and the rest all record verdicts through
these CLI verbs — never by editing
.md renders or the event log directly.
- A rejection must carry concrete, actionable feedback (the
--review-feedback-file); an empty
rejection is not a review.
- Related:
spk-run-review-wp performs the review that produces the verdict; this skill is only the
capture step. spk-run-implement-review drives the loop across WPs.
