verification-before-completion

Require fresh verification command output before claiming work completion.

8|1|Updated Mar 1, 2026

Search Tags:#testing #quality assurance #verification #evidence #ci #commit readiness #prompt guardrails

Ready to run

npx skills add https://github.com/JoeCP17/LLM-Dot-files --skill verification-before-completion-joecp17

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: verification-before-completion
Source: https://github.com/JoeCP17/LLM-Dot-files/tree/main/claude/skills/superpowers/verification-before-completion
Command: npx skills add https://github.com/JoeCP17/LLM-Dot-files --skill verification-before-completion-joecp17

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It prevents false “done/passing” claims by forcing fresh verification evidence before you commit, create a PR, or express satisfaction.

Core Features & Use Cases

Evidence-before-assertions enforcement: Requires running the specific verification command and reading full output (including exit code or failure counts) before any success/completion statement.
Verification gate for common engineering tasks: Applies to tests, linters, builds, bug fixes, regression cycles, and requirement checklists—ensuring you don’t rely on guesses or partial checks.
Anti-rationalization guardrails: Blocks success language (“Great”, “Perfect”, “Done”) and dismissals (“should work”, “agent said so”, “just this once”) unless verification ran in the same message.

Quick Start

Use the verification-before-completion skill right before you commit or open a PR, and include the verification command output that proves the status you want to claim.

Frequently Asked Questions about verification-before-completion

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

FAQPage Schema

How do I prevent false completion claims before opening a pull request?▼

A verification gate prevents false completion claims by identifying the proving command, running it, and checking the exit status and complete output before allowing any success language in PR or commit communications.

What is evidence-based verification in software engineering?▼

Evidence-based verification requires running the specific testing or linting command and reading the complete output, including failure counts, before claiming a task is complete. It blocks success statements unless fresh verification ran in the same message.

How to enforce quality assurance checks before a git commit?▼

To enforce quality assurance checks before a commit, use a verification gate that identifies the proving command, runs it fully, checks the exit status and output, and only then allows completion language in your commit readiness workflow.

Does verification before completion work with linting and CI workflows?▼

Yes, the verification gate applies to common engineering tasks including linting, builds, regression cycles, and CI workflows. It ensures you do not rely on guesses or partial checks by requiring fresh verification evidence before any success statement.

Why are success statements blocked during bug fix regression verification?▼

Success statements are blocked during regression verification to provide anti-rationalization guardrails. This prevents dismissals like should work or agent said so unless the actual verification command ran and its complete output was checked in the same message.

Can I bypass the verification gate for a quick commit just this once?▼

No, the gate includes anti-rationalization guardrails that block success language and dismissals like just this once unless the full verification command ran and its complete output was verified in the same message.