Shokunin verify

Skill vstakhovsky/shokunin-review/.claude/skills/shokunin-verify

Terminal-first validation harness for reviewing PRDs, RFCs, strategy docs, and experiment plans.From the repository description

Install

npx -y skills add vstakhovsky/shokunin-review --skill shokunin-verify

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

1.9 KB, 437 tokens by cl100k_base, as published. Nobody here has run it

Shokunin Verify Skill

Verifies output quality before returning to user.

Purpose

Final verification step to ensure output is safe, grounded, and useful.

When to Use

This skill runs automatically at the end of every review. User doesn't call it directly.

Workflow

Check grounding — Every finding grounded in artifact?
Check actionability — Every finding has concrete fix?
Check specificity — No generic advice?
Check duplicates — No duplicate findings?
Check tone — Tone calm and direct?
Check safety — No toxic language or invented evidence?
Check consistency — Score matches findings?

Inputs

Review findings
Calculated score
Verdict

Output Contract

If verification passes:

Output is returned to user

If verification fails:

Findings revised
Score recalculated if needed
Verification re-run
Only when all checks pass → Output returned

Verification Checklist

✅ All findings reference artifact content
✅ All findings have location (line/section)
✅ All findings have concrete fix
✅ All findings have example (where helpful)
✅ No "improve clarity" without specifics
✅ No "add detail" without guidance
✅ No duplicate findings
✅ Tone is calm, not dramatic
✅ No shaming language
✅ No accusations of AI use
✅ No invented evidence
✅ Score aligns with findings
✅ Verdict matches score

Failure Modes

If check fails:

Finding removed or revised
Output re-verified
Process repeats until all pass

Example

Bad Finding (fails verification):

"Requirements need improvement. Add more detail."

Good Finding (passes verification):

"Requirements section is not testable.
Location: Requirements section
Fix: Rewrite each requirement as Given/When/Then
Example: Given user on checkout, When enters payment, Then order confirmed"

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.