arc-verifying

Enforce evidence-based verification before declaring agent tasks complete.

6|Updated Feb 8, 2026
One-click install
npx skills add https://github.com/GregoryHo/arcforge --skill arc-verifying
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: arc-verifying
Source: https://github.com/GregoryHo/arcforge/tree/main/skills/arc-verifying
Command: npx skills add https://github.com/GregoryHo/arcforge --skill arc-verifying

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill addresses the critical issue of AI agents making completion claims without proper verification, leading to technical debt and unreliable outcomes. It enforces a rigorous verification process before any work is declared complete.

Core Features & Use Cases

  • Enforces Verification: Mandates that all claims of completion are backed by fresh, verifiable evidence.
  • Defines Verification Steps: Outlines a clear "Gate Function" process for identifying, running, reading, and verifying commands.
  • Identifies Common Failures: Lists typical rationalizations and insufficient verification methods.
  • Use Case: Before an agent commits code, it must run the relevant tests and show the output confirming zero failures, preventing the commit of untested code.

Quick Start

Always run the verification command and check its output before claiming any task is complete.

Frequently Asked Questions about arc-verifying

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I prevent AI agents from claiming task completion without verification?

To prevent premature completion claims, enforce a strict verification gate function requiring AI agents to run validation commands and show fresh evidence before declaring work done.

What is a verification gate function for agent workflows?

A verification gate function is a strict process requiring agents to identify, run, read, and verify commands, ensuring all completion claims are backed by fresh, verifiable evidence.

How to verify code quality before an AI agent commits changes?

Verify code quality before committing by running relevant tests and checking the output to confirm zero failures, which enforces a rigorous validation process against untested code.

Why does my AI agent commit code with insufficient verification?

AI agents commit code with insufficient verification due to common rationalizations and skipped validation steps, bypassing the required process of running and reading command outputs.

Can I apply completion checks to all agent tasks?

Yes, you can apply completion checks to all agent tasks requiring a final validation gate, ensuring every declared outcome is validated with fresh evidence before completion.