How does this skill prevent AI hallucinations in coding?

By requiring 'Proof of Work' (logs, command output, and test results), it prevents the AI from stating a task 'should work' without actually verifying it in the environment.

What is 'Anti-Cargo-Cult' verification?

It is a protocol within the skill that requires the AI to explain the 'why' behind a solution, ensuring it isn't just copying patterns without understanding the underlying logic.

What is the Proof of Work skill for Claude Code?

It is a specialized capability that forces the AI to provide technical evidence and verify work through testing before claiming a task is finished.

Does this skill support Test-Driven Development (TDD)?

Yes, it enforces the 'Iron Law' which mandates writing a failing test before any implementation code is written.

Which plugins does Proof of Work depend on?

It primarily depends on the imbue:evidence-logging skill to properly document the validation steps and test outputs it generates.

Proof of Work

Name: Proof of Work
Author: athola

byathola

•

123

•

安全与测试

Enforces rigorous validation, testing, and evidence-based reporting before any task is declared complete.

The Proof of Work skill implements a 'prove before claim' discipline for AI-driven development, preventing premature declarations of success. It mandates a systematic workflow where Claude must reproduce issues, verify solutions with technical evidence, and check for edge cases before providing status updates. By enforcing the 'Iron Law'—no implementation without a failing test—it eliminates cargo-cult programming and ensures that every change is backed by reproducible command outputs and deep technical understanding.

主要功能

01123 GitHub stars

02Iron Law enforcement requiring failing tests before implementation

03Seamless integration with evidence-logging for audit trails

04Mandatory pre-completion validation protocols for all tasks

05Anti-cargo-cult verification to ensure fundamental understanding of code

06Standardized 'Definition of Done' and Acceptance Criteria templates

使用场景

01Maintaining high-integrity development workflows in autonomous coding sessions

02Ensuring bug fixes are validated against reproducible failure cases

03Verifying complex environment configurations and LSP setups

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add athola/claude-night-market proof-of-work

For use in Claude.ai and ChatGPT

Download Skill