How does it handle regression testing?

It mandates a Red-Green cycle where the agent must prove a test fails without the fix and passes with it, ensuring the test is actually targeting the specific issue.

What is the main purpose of the Verification Before Completion skill?

It ensures that any claim of success or task completion is backed by fresh, verifiable evidence from command outputs rather than assumptions or 'seems correct' statements.

Does this skill work with any testing framework?

Yes, it is framework-agnostic and applies the principle of evidence-before-assertion to any command-line tool, including linters, compilers, and test suites.

When does this skill trigger during a workflow?

It applies before any commit, pull request creation, or any statement implying that a task is finished, 'looks good,' or is 'successfully fixed.'

Why should I use this skill instead of standard prompts?

This skill provides a structured 'Gate Function' that prevents common AI hallucinations regarding task completion and enforces a rigorous check against actual terminal output.

Verification Before Completion

Name: Verification Before Completion
Author: ZhanlinCui

byZhanlinCui

•

115

•

Security & Testing

Enforces a strict evidence-based workflow requiring verifiable proof through command execution before any task is declared finished or successful.

The Verification Before Completion skill implements a rigorous quality assurance protocol for Claude by mandating fresh verification evidence before any completion claims are made. It eliminates guesswork and false confidence by requiring the agent to identify necessary verification commands, execute them fully, and analyze the output against specific requirements before committing code or opening PRs. This 'Gate Function' approach prevents common pitfalls such as broken builds, unverified bug fixes, and partial linter checks, ensuring that every assertion of success is backed by actual technical evidence rather than assumptions.

Key Features

01Evidence-based assertion requirements for all success claims

02Mandatory Gate Function protocol for systematic verification

03115 GitHub stars

04Red-Green TDD cycle verification for robust regression testing

05Strict identification and prevention of 'false confidence' language

06Independent verification of delegated agent tasks and VCS diffs

Use Cases

01Preventing broken builds by verifying compilation status before PR creation

02Ensuring bug fixes are valid through rigorous Red-Green regression cycles

03Validating complex requirement checklists against actual output before task handoff

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add zhanlincui/agent-skills-hunter verification-before-completion

For use in Claude.ai and ChatGPT

Download Skill