# Classification Validate

> Validate Kokoro experiment and listening-decision labels, gate categories, and result manifests. Use for pass/warn/fail decision records, F0 or audio listening reviews, bakeoff classifications, or preventing invented result categories. Do not use for model export numerical parity or generic data labeling.

- Skill: `mattmireles/classification-validate` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add mattmireles/classification-validate`
- Raw SKILL.md: https://api.skillmd.com/api/skills/mattmireles/classification-validate/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: mattmireles (https://skillmd.com/u/mattmireles)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/mattmireles/classification-validate

---


# Validate Kokoro Decision Labels

1. Read [the repository anchors](references/index.md).
2. Identify the canonical schema or allowed decision values from the script,
   plan, and existing validated artifact.
3. Reject missing or unknown values; do not normalize a label into a new
   category after observing results.
4. Separate hard failures, caveats requiring evidence, and provisional
   observations. Preserve the underlying measurements and reviewer notes.
5. Rerun the owning validator before treating a summary or bakeoff table as
   complete.

## Guardrails

- Do not turn a subjective listening decision into numerical parity evidence.
- Do not rewrite a negative result into a warning to make a gate pass.
- Keep source audio, model revision, device, and artifact identity attached to
  every decision record.