Does it only check automated tests?

No, it also covers manual verification criteria for visual UI changes, accessibility, and runtime behaviors that automated tests might miss.

Which tech stacks are supported?

The skill includes pre-defined verification commands for React (Vite), Next.js, Node/Express, Expo, Prisma, and Monorepos using Turbo.

What does the Verification Guard skill do?

It mandates that Claude runs specific terminal commands, such as tests, builds, or linting, and verifies the actual output before it can claim a task is finished.

How does it prevent 'hallucinated' success?

It requires a 'proof command' to be executed and the terminal output to be read, preventing the AI from assuming code works without execution evidence.

Verification Guard

Name: Verification Guard
Author: hjemmesidekongen

byhjemmesidekongen

0•

安全与测试

Enforces a mandatory verification protocol using stack-specific commands to ensure tasks meet quality and acceptance criteria before completion.

Verification Guard acts as a rigorous quality assurance gate for Claude, preventing it from marking tasks as complete without empirical evidence. It provides a structured framework for running tests, builds, type-checking, and linting across modern tech stacks like React, Next.js, and Node.js. By requiring the model to execute proof commands and verify edge cases, this skill eliminates common AI pitfalls such as hallucinated success, broken imports, and silent type failures, ensuring that all output is genuinely production-ready.

主要功能

010 GitHub stars

02Edge case and error path verification protocol

03Automated type-checking (tsc) and linting validation

04Mandatory verification checkpoints for tests, builds, and linting

05Manual review checklists for UI/UX and accessibility

06Stack-specific command references for React, Next.js, Node, and Prisma

使用场景

01Preventing broken builds from being committed by enforcing local build checks

02Ensuring full type safety in TypeScript projects beyond standard build stripping

03Verifying complex edge cases and error handling before finalizing features

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add hjemmesidekongen/ai verification-before-completion

For use in Claude.ai and ChatGPT

主要功能

010 GitHub stars

02Edge case and error path verification protocol

03Automated type-checking (tsc) and linting validation

04Mandatory verification checkpoints for tests, builds, and linting

05Manual review checklists for UI/UX and accessibility

06Stack-specific command references for React, Next.js, Node, and Prisma

使用场景

01Preventing broken builds from being committed by enforcing local build checks

02Ensuring full type safety in TypeScript projects beyond standard build stripping

03Verifying complex edge cases and error handling before finalizing features

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add hjemmesidekongen/ai verification-before-completion

For use in Claude.ai and ChatGPT