skills.rest
ananddtyagi
ananddtyagiCOMMUNITY
Explore all skills by this author →
Website𝕏@ananddtyagiGitHub

AI Truthfulness Enforcer

Validates AI success claims against live build tests, browser evidence, and user confirmation.

688|85|Updated Oct 9, 2025
Search Tags:#verification#evidence#playwright#Testing#Debugging#claim validation#Code quality
One-click install
npx skills add https://github.com/ananddtyagi/cc-marketplace --skill ai-truthfulness-enforcer
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: AI Truthfulness Enforcer
Source: https://github.com/ananddtyagi/cc-marketplace/tree/main/plugins/claude-dev-infrastructure/skills/ai-truthfulness-enforcer
Command: npx skills add https://github.com/ananddtyagi/cc-marketplace --skill ai-truthfulness-enforcer

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

AI coding assistants often report tasks as "fixed", "working", or "complete" without real verification, eroding trust and shipping broken code. This Skill intercepts success claims and forces evidence-based validation before any such statement is allowed.

Core Features & Use Cases

  • Claim Detection & Interception: Automatically detects trigger phrases like "works", "fixed", or "X% complete" and halts output until evidence is provided.
  • Mandatory Evidence Collection: Requires live build runs, Playwright browser tests, timestamped screenshots, console error monitoring, and error-count verification for every claim.
  • User Verification Gate: Blocks all fix claims until the user explicitly confirms via an AskUserQuestion prompt that the issue is actually resolved.
  • Use Case: After fixing a Vue.js bug, the assistant runs npm run build and npx vue-tsc --noEmit, captures screenshots, then asks you to confirm the fix before marking the task complete.

Quick Start

Ask the assistant to fix a bug in your project and it will run real tests, capture evidence, and request your confirmation before claiming the fix works.

Frequently Asked Questions about AI Truthfulness Enforcer

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

FAQPage Schema
How do I stop Claude Code from making false success claims?▼

Install a verification skill that intercepts claim phrases like "fixed" or "works" and demands live evidence. It requires real build runs, browser tests, and screenshots before any success statement is allowed.

How to verify AI bug fixes before marking tasks complete?▼

Require the assistant to run builds and type checks, capture timestamped screenshots, monitor the console for errors, then explicitly ask you to confirm the fix. The task stays open until you verify it works.

What evidence is required for build status claims?▼

Build claims require a live `npm run build` execution, full console output capture, the actual success or failure message, measured build time, and a timestamped terminal screenshot. Claims without this evidence are automatically rejected.

Does this verification work with Playwright browser testing?▼

Yes, functionality claims require live browser sessions via Playwright MCP, including before and after screenshots, zero console errors, cross-view testing, and data persistence checks after page refresh.

What happens when a false claim is detected?▼

Detection triggers an immediate halt of all work, an audit of recent claims, and a lockdown preventing further claims until verification. Repeated offenses escalate to mandatory verification for all statements.

Related Skills

Explore More Software Engineering Skills→
spencerpaulyspencerpauly
672

adding-e2e-tests

Set up Playwright end-to-end testing with smoke tests and GitHub Actions.

Community
Intermediate
coff33ninjacoff33ninja

release-changelog

Automates SemVer-Compliant release workflows including bumping versions, updating changelogs, and triggering CI/CD pipelines for diverse programming languages and systems.

Community
Advanced
FilenCloudDiensteFilenCloudDienste
87

verify-changes

Execute ESLint, TypeScript checks, and unit tests on code changes.

Official
Intermediate
tody-agenttody-agent
48

cm-quality-gate

Enforce test gates and evidence-based verification before deployment in CI/CD workflows.

Community
Advanced