verification

Run verification commands and inspect exit codes before claiming completion.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/ForeverWorld/curdx-ralph --skill verification-foreverworld
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: verification
Source: https://github.com/ForeverWorld/curdx-ralph/tree/main/skills/verification
Command: npx skills add https://github.com/ForeverWorld/curdx-ralph --skill verification-foreverworld

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Prevents premature or false claims of completion by enforcing an evidence-first verification process that requires running the actual verification command and inspecting its output and exit code before asserting success.

Core Features & Use Cases

  • Evidence-first gate: A clear five-step gate to identify the proof command, run it fresh, read full output, verify results, and only then claim success.
  • Common applications: Use before committing, creating PRs, moving tasks to done, delegating to agents, or reporting test/lint/build status.
  • Failure prevention: Reduces regressions, shipping incomplete features, and trust erosion by requiring reproducible verification evidence.

Quick Start

Ask verification to run the relevant test or build command, include the full output and exit code, and then state whether the claim is supported by evidence.

Frequently Asked Questions about verification

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

FAQPage Schema
How do I enforce verification checks before claiming a task is complete?

To enforce verification before claiming completion, use an evidence-first process that runs the actual verification command, reads the exit code and full output, and presents evidence before asserting success.

Why does my CI pipeline report false build success without running tests?

False build success occurs when claims lack fresh verification evidence. Running the full test command and inspecting its complete output prevents premature completion claims and reduces regressions.

What is the best way to verify code review outputs before creating a pull request?

The best way to verify outputs before a pull request is applying a five-step gate: identify the proof command, execute it, read full output, verify results, and only then claim success.

Can I use this verification gate for task transitions and delegating results?

Yes, this verification gate applies when preparing commits, moving tasks to done, delegating to agents, or reporting test and build status by requiring reproducible verification evidence.

How do I prevent shipping incomplete features during software development workflows?

Prevent shipping incomplete features by requiring an evidence-first verification process that identifies the proof command, runs it fresh, and reads exit codes before asserting success.