completion-gate

Validate agent completion claims against command output, diffs, and file evidence.

1|Updated May 18, 2026

Search Tags:#validation #evidence #quality gate #claims #build verification #stub detection #test gating

Ready to run

npx skills add https://github.com/linenoize/topia --skill completion-gate-linenoize

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: completion-gate
Source: https://github.com/linenoize/topia/tree/main/skills/completion-gate
Command: npx skills add https://github.com/linenoize/topia --skill completion-gate-linenoize

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill prevents agents from claiming work is finished without proof, reducing false completions, unverified test results, and stubbed implementations.

Core Features & Use Cases

Evidence validation: Checks completion claims against tool output, diffs, and file evidence before the work is accepted.
Stub and gap detection: Flags placeholder code, missing verification, and incomplete changes that only appear finished.
Use Case: Use this after a build, fix, or merge task when an agent says everything is complete and you need proof that the result is real.

Quick Start

Use the completion-gate skill to verify the agent’s completion claims against the latest evidence and report any missing or contradictory proof.

Frequently Asked Questions about completion-gate

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

FAQPage Schema

How do I verify agent completion claims against actual build or test output?▼

You validate completion claims by matching them against command output, diffs, and self-validation checks to produce confirmed, unconfirmed, or contradicted verdicts. This blocks false done reports by requiring proof for build, test, fix, and merge workflows.

What is stub detection in code change verification and when do I need it?▼

Stub detection flags placeholder code and incomplete changes that only appear finished. You need this quality gate when an agent reports a task is complete but the implementation may contain gaps or unverified test results requiring proof.

How do I stop false done reports from automated coding agents?▼

Stop false done reports by applying a quality gate that requires matching agent claims to command output, diffs, and file stubs. This blocks unconfirmed or contradicted verdicts, preventing agents from accepting work without proof.

Can I use a quality gate to check for missing verification in merge workflows?▼

Yes, you can apply a quality gate to check for missing verification in merge workflows. It validates completion claims against diffs and test runs, producing contradicted verdicts if code changes lack matching evidence before acceptance.

What are the limitations of evidence validation for build verification?▼

Evidence validation for build verification is limited to checking claims against command output, diffs, and file stubs. It produces unconfirmed verdicts when proof is missing but cannot execute new tests or fix the underlying code gaps it detects.