Why is the Verification Guardrail skill important for AI coding?

It prevents AI agents from making assumptions that code works without actually running the verification commands, ensuring high-trust development.

What happens if a linter passes but the build fails?

The skill treats these as distinct requirements; a linter passing is not sufficient evidence to claim the build succeeds.

What is the 'Iron Law' in this skill?

The Iron Law dictates that no completion claims can be made without fresh verification evidence generated in the current session.

Does it support regression testing?

Yes, it specifically encourages a red-green cycle where the AI must prove a test fails before proving that its fix makes the test pass.

Can this skill be bypassed for simple fixes?

No, the protocol is designed to be non-negotiable to maintain consistent quality and prevent the 'just this once' fallacy.

Verification Guardrail

Name: Verification Guardrail
Author: julianromli

byjulianromli

•

101

•

Security & Testing

Enforces rigorous evidence-based verification protocols before any task is marked as complete or successful.

The Verification Guardrail skill implements a strict 'evidence-before-claims' policy for Claude Code, preventing premature success declarations or false positives during the development lifecycle. By mandating fresh verification runs, output analysis, and red-green testing cycles, it ensures that every assertion—from passing tests to bug fixes—is backed by real-time data rather than assumptions. This skill is essential for maintaining high trust and code quality, effectively eliminating the risk of 'should work' assertions in collaborative environments.

Key Features

01Rationalization Prevention: Automatically flags and stops the use of non-committal language like 'should' or 'probably'.

02Evidence Checklist: Maps specific claims (Build, Lint, Test) to their required command-line output counterparts.

03The Iron Law Protocol: Blocks any success claim that lacks fresh, command-based verification evidence.

04Systematic Gate Function: Guides the AI through a mandatory Identify-Run-Read-Verify sequence.

05101 GitHub stars

06Red-Green Regression Checks: Mandates verifying the failure state before confirming a fix to prevent false positives.

Use Cases

01Ensuring a reported bug is truly resolved through systematic red-green regression testing cycles.

02Verifying unit and integration tests pass before committing code to a repository to maintain build integrity.

03Validating full build and linter status before creating a pull request to minimize CI/CD failures.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add julianromli/opencode-template verification-before-completion

For use in Claude.ai and ChatGPT

Download Skill