reviewing-completed-work

Name: reviewing-completed-work
Availability: InStock
Author: michaelkd01

Validate completed tasks against acceptance criteria and test contracts.

Updated Aug 27, 2026

Search Tags:#drift-detection #claude-code #acceptance-criteria #paperclip #supervisor-validation #execution-verdicts #testing-checklist

Ready to run

npx skills add https://github.com/michaelkd01/dic-skills --skill reviewing-completed-work

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: reviewing-completed-work
Source: https://github.com/michaelkd01/dic-skills/tree/main/reviewing-completed-work
Command: npx skills add https://github.com/michaelkd01/dic-skills --skill reviewing-completed-work

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill enforces rigorous supervisor validation so test specs and execution outputs never drift from the acceptance criteria or documented architecture before issuing a verdict.

Core Features & Use Cases

Test Spec Mode reviews Step 3.5 artifacts, checks against CLAUDE.md conventions, and flags missing coverage, trivial checks, or scope creep before execution.
Execution Mode inspects completed Claude Code runs, compares evidence with each acceptance criterion, confirms the approved test contract, and watches for architectural drift or scope violations.
Decision Protocols force explicit PASS/FIX/REJECT determinations with documented evidence, ensuring assignments receive the correct follow-up, status updates, and deploy actions.

Quick Start

Validate the latest execution output, confirm each AC and test contract, then publish the verdict as PASS, FIX, or REJECT.

Frequently Asked Questions about reviewing-completed-work

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

FAQPage Schema

How do I validate completed Claude Code runs against acceptance criteria?▼

Validate completed Claude Code runs by inspecting execution outputs, comparing evidence against each acceptance criterion, and confirming the approved test contract to issue a final verdict.

What is drift detection in software execution validation?▼

Drift detection in execution validation identifies architectural drift or scope violations by comparing completed Claude Code run evidence against documented conventions in CLAUDE.md and Notion architecture docs.

How do I review test specs for missing coverage before execution?▼

Review test specs by checking Step 3.5 artifacts against CLAUDE.md conventions to flag missing coverage, trivial checks, or scope creep before execution begins.

When should I issue a FIX versus REJECT verdict for a completed task?▼

Issue a FIX verdict when execution outputs require corrections to meet acceptance criteria, and issue a REJECT when outputs fundamentally violate documented architecture or scope, ensuring documented evidence supports the decision.

Does this supervisor validation workflow work with Paperclip missions?▼

Yes, this reviewer workflow applies exclusively to Paperclip supervisor missions, covering both test spec reviews and execution checks after Claude Code runs to validate project conventions.

What are the limitations of automated acceptance criteria validation?▼

Automated acceptance criteria validation requires confirming project conventions from CLAUDE.md and Notion architecture docs before issuing verdicts, and cannot bypass the explicit PASS, FIX, or REJECT decision protocol.