skills.rest
nthpaul
nthpaulCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

verify-this

Compare baseline and treatment artifacts to verify falsifiable claims.

Updated May 18, 2025
Search Tags:#performance#debugging#verification#evidence#experimentation#api-diff#ui-behavior
One-click install
npx skills add https://github.com/nthpaul/dotfiles --skill verify-this-nthpaul
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: verify-this
Source: https://github.com/nthpaul/dotfiles/tree/main/cursor/.cursor/plugins/cache/cursor-public/cursor-team-kit/d1cdb88a9eb33cf392395c87e3fd76419fc1010e/skills/verify-this
Command: npx skills add https://github.com/nthpaul/dotfiles --skill verify-this-nthpaul

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill removes guesswork by turning a claim into repeatable verification with concrete baseline vs treatment evidence instead of relying on recap or opinions.

Core Features & Use Cases

  • Falsifiable Restatement: Converts the user’s claim into a measurable, disprovable statement with a metric and threshold.
  • Baseline vs Treatment Evidence: Captures comparable artifacts before and after the change (or between states) using the same command, data, warmup, and environment.
  • Clear Verdict Output: Produces exactly one of VERIFIED, NOT VERIFIED, or INCONCLUSIVE based on whether the evidence matches the predicted direction and threshold.
  • Common Scenarios: Validates bug fixes with before/after repros, confirms UI/CLI/API behavior with transcripts or screenshots, and supports performance and memory claims with timings, profiles, or heap snapshots.

Quick Start

Ask to verify a specific claim by providing the falsifiable metric/threshold and the baseline-to-treatment command or repro context, then request a verdict of VERIFIED, NOT VERIFIED, or INCONCLUSIVE with the evidence artifacts summarized.

Frequently Asked Questions about verify-this

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

FAQPage Schema
How do I verify a bug fix using before and after evidence?▼

Verifying a bug fix requires capturing baseline and treatment evidence using the same command and environment, then comparing artifacts to produce a VERIFIED, NOT VERIFIED, or INCONCLUSIVE verdict.

What is the best way to prove a performance claim with concrete data?▼

Proving a performance claim involves restating it as a falsifiable metric with a threshold, capturing baseline vs treatment timings or profiles, and comparing the artifacts to confirm the predicted improvement direction.

Can I validate API request and response changes using baseline and treatment diffs?▼

Validating API behavior changes uses controlled baseline and treatment captures of request and response artifacts, comparing the diffs to determine if the observed behavior matches the predicted claim outcome.

How do I structure a falsifiable claim for debugging and change validation?▼

Structuring a falsifiable claim requires defining a measurable metric and a specific threshold, enabling controlled baseline and treatment captures that produce comparable evidence for a constrained verdict.

Does this approach work for checking CLI and TUI behavior changes?▼

Checking CLI and TUI behavior changes works by capturing baseline and treatment transcripts or screenshots under identical conditions, then comparing the artifacts to reach a clear verification verdict.

Why do I get an INCONCLUSIVE verdict when comparing baseline and treatment evidence?▼

An INCONCLUSIVE verdict occurs during evidence comparison when the baseline and treatment artifacts lack sufficient clarity or repeatability to definitively confirm or deny the predicted claim threshold.

Related Skills

Explore More Software Engineering Skills→
ksciuskscius
2

verify-this

Validate claims by comparing baseline and treatment artifacts into a verdict.

Community
Advanced
MihaiOnSoftwareMihaiOnSoftware
1

write-validation-report

Aggregate validation results into a dated pass/fail Markdown report.

Community
Basic
cursorcursor
2.6k

verify-this

Compare baseline and treatment artifacts to emit a verification verdict.

Official
Advanced
CDO-07-NewCDO-07-New

eval-harness

Define repeatable evaluation harnesses for Claude Code task outcomes.

Official
Intermediate