verification-before-completion

Name: verification-before-completion
Availability: InStock
Author: adamatdevops

Run full verification commands and check exit codes before reporting task completion.

1|Updated Jan 7, 2026

Search Tags:#verification #quality gates #test evidence #agent workflows #completion claims #exit code checking #honest reporting

Ready to run

npx skills add https://github.com/adamatdevops/forge-works --skill verification-before-completion-adamatdevops

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: verification-before-completion
Source: https://github.com/adamatdevops/forge-works/tree/main/.skills/obra/verification-before-completion
Command: npx skills add https://github.com/adamatdevops/forge-works --skill verification-before-completion-adamatdevops

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It prevents dishonest or premature “done” claims by forcing verification evidence to be collected and checked before reporting success, satisfaction, or task completion.

Core Features & Use Cases

Evidence-first completion rule: Requires fresh verification output for any completion/status claim, including PR/commit readiness and “finished” messaging.
Gate function workflow: Guides you to identify the proving command, run the full command, read outputs/exit codes, confirm with evidence, and only then claim.
Red-flag and failure-mode avoidance: Stops common rationalizations like “should pass,” “looks correct,” trusting agent reports, partial checks, or counting earlier runs.

Quick Start

Use the verification-before-completion skill to ensure you do not claim tests/builds/lints succeeded until you run the full verification command and report what the output proves.

Frequently Asked Questions about verification-before-completion

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

FAQPage Schema

How do I stop my AI agent from falsely claiming a task is done before tests pass?▼

To stop false completion claims, you must enforce a verification gate that requires fresh test evidence before reporting success. This prevents premature done claims by checking exit codes and outputs to ensure the proving command actually passed.

What is evidence-first completion reporting in software engineering workflows?▼

Evidence-first completion reporting is a workflow rule requiring fresh verification output, such as passing test or lint results, before claiming a task is finished. It forces you to run the full proving command and confirm the output proves the stated result.

How do I verify bug fixes and prevent premature success claims in CI pipelines?▼

You can verify bug fixes by applying a gate function workflow that runs the complete verifying command, checks the exit code, and reads the output. This prevents premature success claims by allowing completion reporting only when evidence explicitly confirms the result.

Why does my agent workflow report task completion when the build or lint fails?▼

Agent workflows report completion incorrectly when they rely on rationalizations like should pass or trusting previous runs instead of fresh evidence. To fix this, enforce an exit code checking rule that stops honest reporting until the full verifying command is run and confirmed.

Does this verification gate workflow work for regression testing and linting?▼

Yes, this verification gate workflow applies directly to regression testing, linting, and builds within software engineering. It ensures you do not claim tests or lints succeeded until you run the full verification command and report what the output proves.

What are common failure modes when reporting task completion in automated workflows?▼

Common failure modes include trusting agent reports without fresh checks, relying on partial checks, or counting earlier runs as proof. Avoid these red flags by enforcing a gate workflow that runs the complete verifying command and checks the exit code before reporting.