Why is this skill important for AI-driven development?

It prevents AI agents from reporting 'success' based on assumptions, ensuring that every change is backed by actual terminal output and exit codes.

What is the core principle of this Claude Code skill?

The core principle is 'Evidence before claims, always.' It prevents Claude from claiming work is done until it has run the necessary commands to prove it.

How does the 'Gate Function' work?

Before making any success claim, the skill requires Claude to identify the correct command, run it, read the full output, and verify that the output confirms the claim.

When should the verification process be applied?

It must be applied before any positive statement about work state, before committing code, creating pull requests, or moving to a new task.

Can this skill help with regression testing?

Yes, it specifically enforces a Red-Green cycle where a test must be shown to fail before the fix is applied and then shown to pass afterward.

Verification Before Completion

Name: Verification Before Completion
Author: SwervoJake

bySwervoJake

0•

Security & Testing

Enforces a rigorous evidence-first workflow by requiring successful command execution before any claims of task completion or code correctness.

The Verification Before Completion skill implements a non-negotiable gate function for Claude Code, ensuring that no work is declared finished, fixed, or passing without fresh, empirical evidence. By mandating the identification and execution of specific verification commands—such as test suites, linters, or build scripts—before any success assertions are made, it eliminates the 'should work' fallacy and prevents the introduction of regressions. This skill is essential for developers who demand high-integrity code and want to eliminate false positives in their automated or manual workflows.

Key Features

01Integrated TDD Red-Green cycle verification for robust regression testing

02Independent verification protocols for checking delegated AI agent outputs

03Mandatory 'Gate Function' requiring evidence before any success assertions

040 GitHub stars

05Standardized verification patterns for tests, builds, linting, and bug fixes

06Prevention of common rationalizations like 'it should work' or 'I'm confident'

Use Cases

01Validating that complex multi-step requirements are fully met via line-by-line checklists

02Ensuring build integrity and linter compliance before committing code to a repository

03Verifying bug fixes through documented regression test failure and subsequent passing

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add swervojake/claaaude-studio verification-before-completion

For use in Claude.ai and ChatGPT

Key Features

01Integrated TDD Red-Green cycle verification for robust regression testing

02Independent verification protocols for checking delegated AI agent outputs

03Mandatory 'Gate Function' requiring evidence before any success assertions

040 GitHub stars

05Standardized verification patterns for tests, builds, linting, and bug fixes

06Prevention of common rationalizations like 'it should work' or 'I'm confident'

Use Cases

01Validating that complex multi-step requirements are fully met via line-by-line checklists

02Ensuring build integrity and linter compliance before committing code to a repository

03Verifying bug fixes through documented regression test failure and subsequent passing

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add swervojake/claaaude-studio verification-before-completion

For use in Claude.ai and ChatGPT