Does it support multiple programming languages?

Yes, it is language-agnostic and provides command patterns for JavaScript, TypeScript, Python, and other major ecosystems using their native test and build tools.

How does this skill prevent AI hallucinations?

It mandates executable proof, such as specific exit codes and command outputs, before a task is considered finished, ensuring the AI cannot claim code works without actually running tests.

Can it be used for deployment verification?

Yes, it includes specific protocols for checking deployment status, environment health, and verifying rollback capabilities in production or staging environments.

What are the minimum requirements for verification?

At a minimum, the skill requires at least one verification type executed, an exit code of 0 captured, a timestamp recorded, and the evidence stored in the project context.

What is the 'Gold Standard' for evidence?

The Gold Standard includes 80% code coverage, no linter warnings, performance benchmarks within thresholds, and successful accessibility audits.

Evidence Verification

Name: Evidence Verification
Author: yonatangross

byyonatangross

•

Security & Testing

Verifies task completion by collecting and documenting executable proof including test results, build outputs, and exit codes.

The Evidence Verification skill ensures that AI-generated code and tasks meet production standards by enforcing a 'show, don't tell' policy for all work. It automates the collection of test suite results, coverage metrics, build artifacts, and linter outputs, preventing hallucinations and ensuring that no task is marked complete without verifiable evidence. By integrating standardized protocols for verification, this skill builds confidence during agent handoffs, pull request creation, and production deployments, making it an essential component for high-reliability AI development workflows.

Key Features

01Standardized templates for test, build, and code quality reporting

02Automated exit code validation to confirm command success

03Production-ready deployment and health check verification protocols

04Tiered quality standards from minimum viable to gold-standard requirements

05Seamless integration with session state for persistent evidence storage

0669 GitHub stars

Use Cases

01Ensuring build stability and security compliance before production deployment

02Validating code implementations before marking tasks as complete

03Conducting automated quality assurance checks during code reviews

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add yonatangross/orchestkit evidence-verification

For use in Claude.ai and ChatGPT

Download Skill