verify-this

Run baseline and treatment comparisons to falsify user claims and output a verdict.

Updated May 22, 2026
One-click install
npx skills add https://github.com/paulbrittain/claude-team-kit --skill verify-this-paulbrittain
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: verify-this
Source: https://github.com/paulbrittain/claude-team-kit/tree/main/skills/verify-this
Command: npx skills add https://github.com/paulbrittain/claude-team-kit --skill verify-this-paulbrittain

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill replaces vague “trust me” answers with repeatable, falsifiable verification so you can determine whether a specific claim is actually true under comparable baseline and treatment conditions.

Core Features & Use Cases

  • Falsifiable claim restatement: Converts the user’s request into a measurable condition, metric, and threshold that can be disproved.
  • Baseline vs treatment evidence capture: Collects comparable artifacts from the old state and the changed state using the same command, data, warmup, and environment.
  • Clear, binary outcome: Produces exactly one verdict—VERIFIED, NOT VERIFIED, or INCONCLUSIVE—with named evidence and concise reasoning tied to confounds.

Quick Start

Ask the AI to verify your claim by running a baseline and treatment comparison and returning VERIFIED, NOT VERIFIED, or INCONCLUSIVE with the exact evidence deltas.

Frequently Asked Questions about verify-this

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

FAQPage Schema
How do I verify code behavior changes with baseline vs treatment comparison?

Baseline vs treatment comparison verifies code behavior changes by collecting comparable artifacts from the old state and changed state using the same command, data, warmup, and environment to accurately measure evidence deltas.

Can I capture UI screenshots and CLI output for CI troubleshooting?

Yes, you can capture UI screenshots and CLI output for CI troubleshooting by applying matched measurement surfaces to collect and compare local testing artifacts from both baseline and treatment states.

What is a falsifiable claim restatement for performance benchmarking?

Falsifiable claim restatement for performance benchmarking converts a user-provided request into a measurable condition, metric, and threshold that can be disproved under comparable local testing conditions.

Does API diff verification require specific dependencies or components?

API diff verification requires no specific dependencies or components, relying solely on restated falsifiable criteria, matched measurement surfaces, and artifact capture to produce a definitive verdict.

What verdicts can I expect from test evidence verification?

Test evidence verification outputs exactly one verdict among VERIFIED, NOT VERIFIED, or INCONCLUSIVE, providing named evidence and concise reasoning tied to confounds for each result.

Why does my baseline vs treatment comparison return an INCONCLUSIVE verdict?

A baseline vs treatment comparison returns an INCONCLUSIVE verdict when confounds prevent a definitive falsification, meaning the collected evidence could not conclusively prove or disprove the stated threshold.