verify-before-done

Enforce a three-level evidence rubric for verifying completion claims.

8|1|Updated Apr 10, 2026
One-click install
npx skills add https://github.com/max4c/skills --skill verify-before-done-max4c
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: verify-before-done
Source: https://github.com/max4c/skills/tree/main/skills/verify-before-done
Command: npx skills add https://github.com/max4c/skills --skill verify-before-done-max4c

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The discipline of proving work is done before claiming it's done. Evidence first, assertions second. Every success claim must be backed by an observable signal.

Core Features & Use Cases

  • Three levels of evidence (Mechanical, Behavioral, Consensus) guide verification to different degrees of certainty.
  • Red-flag detection helps surface hedges like "should" or "looks good" and prompts explicit validation.
  • Triggers during reviews or status changes (e.g., Ready to merge, Deploy it) ensure claims are backed by observable signals.

Quick Start

Apply the verification framework to the item you're about to mark as done and document the observable signals.

Frequently Asked Questions about verify-before-done

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I enforce evidence-based verification before closing a software development ticket?

Evidence-based verification is enforced by applying a three-level rubric (Mechanical, Behavioral, Consensus) that requires explicit named claims and documented run instructions to produce observable signals before ticket closure. This prevents unverified success assertions.

What is the best way to prove code changes are ready to merge?

The best way to prove code changes are ready to merge is to trigger a structured verification process that captures observable signals across mechanical, behavioral, and consensus levels. This validates claims with concrete evidence rather than assumptions.

How does end-to-end quality assurance verification handle vague claims like looks good?

End-to-end quality assurance verification handles vague claims through red-flag detection, which surfaces hedges like should or looks good and prompts explicit validation. This ensures every success claim is backed by an observable signal.

Can I use a structured verification rubric for deployment status changes?

Yes, you can use a structured verification rubric for deployment status changes. It applies during status transitions like ready to merge or deploy it, ensuring that all success claims are backed by observable signals before completion.

What are the limitations of using a three-level evidence rubric for code review?

A limitation of using a three-level evidence rubric for code review is that it requires strict discipline to document run instructions and explicit named claims. Teams must consistently apply mechanical, behavioral, and consensus checks to avoid incomplete verification.