Does the verify skill decide if a task is successfully finished?

No, the verify skill is responsible for running tests and producing result files. The 'check' skill is used to evaluate those results and render a final verdict.

Can this skill be used for tasks other than writing code?

Yes, it is domain-adaptive. It uses type profiles to determine appropriate verification standards for various tasks like documentation, design, or research.

How does verify integrate with Git?

It automatically commits verification results to the specific task branch with standardized commit messages, ensuring a transparent history of test results.

What is the difference between quick and full checkpoints?

A 'quick' checkpoint focuses on fast feedback like build, lint, and type checks. A 'full' checkpoint executes the comprehensive suite including acceptance and regression tests.

What are VFP Metrics in the verification report?

For software tasks, it tracks Verify-First Protocol metrics such as stub counts, green/red status, and compliance with VH→HS (Verify-Hypothesize) disciplines.

Task-AI: Test Verification

Name: Task-AI: Test Verification
Author: huacheng

byhuacheng

•

安全与测试

Runs domain-adapted tests and verification procedures to provide structured evidence for task completion and quality.

The Task-AI: Verify skill is a specialized component of the moonview ecosystem designed to handle sophisticated test execution across different development stages. It automates the running of linting, type checks, unit tests, and acceptance criteria based on specific task profiles. Unlike standard test runners, it intelligently adapts its strategy using domain standards and historical experience, producing detailed markdown reports and VFP (Verify-First Protocol) metrics. It separates the execution of tests from the rendering of verdicts, allowing for a robust, auditable verification trail that integrates directly into the AI-driven development lifecycle.

主要功能

01Domain-adaptive testing strategies tailored to specific task types and profiles

02Automated git-based versioning and structured reporting of all test outcomes

03Integrated Verify-First Protocol (VFP) metrics for software development tracking

04Automated generation of structured test templates for skill and module development

051 GitHub stars

06Multi-checkpoint support including quick linting, full suites, and step-specific verification

使用场景

01Automating fast feedback loops via build and lint checks during active execution steps

02Generating standardized test scaffolds for new AI skills and domain-specific modules

03Performing comprehensive post-execution verification against target acceptance criteria

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add huacheng/moonview verify

For use in Claude.ai and ChatGPT

Download Skill