review-feature

Audit implemented feature behavior against documented specifications without modifying code.

Updated Jun 3, 2026
One-click install
npx skills add https://github.com/0xdeafcafe/skills --skill review-feature
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: review-feature
Source: https://github.com/0xdeafcafe/skills/tree/main/skills/review-feature
Command: npx skills add https://github.com/0xdeafcafe/skills --skill review-feature

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill helps you verify whether a feature is actually complete, correct, and aligned with its intended behavior before you merge it. It focuses on end-to-end feature logic, surfacing missing edge-case handling, broken flows, undocumented side effects, and mismatches between the implementation and the spec.

Core Features & Use Cases

  • Spec-first feature auditing: Finds ADRs, specs, and PR context before reviewing code so the audit is grounded in documented intent.
  • End-to-end flow tracing: Maps the feature from entry points through services, repositories, external calls, and side effects to catch logic gaps across layers.
  • Structured findings output: Reports issues in a consistent finding format for follow-up workflows such as fix orchestration or broader PR review.
  • Use Case: Review a pull request that adds order cancellation and confirm that validation, retries, refunds, emails, loading states, and error responses all match the promised behavior.

Quick Start

Ask the assistant to review the feature in the current PR for completeness against its ADR or spec and return findings only.

Frequently Asked Questions about review-feature

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

FAQPage Schema
How do I audit a pull request for feature completeness against a spec?

Auditing a pull request for feature completeness involves discovering ADRs and specs, tracing end-to-end logic across UI and persistence layers, and emitting structured findings for missing edge cases or side effects without modifying code.

What is spec-first feature auditing in code review?

Spec-first feature auditing grounds code review in documented intent by locating ADRs and specifications before evaluating implementation behavior, ensuring validation checks, error handling, and loading states match the promised feature logic.

How do I find missing edge cases and error handling in a GitHub pull request?

Finding missing edge cases and error handling requires mapping the feature from entry points through services to external calls, validating spec compliance across API and persistence layers, and reporting broken flows as structured findings.

Can I validate end-to-end feature logic across multiple layers without modifying code?

Validating end-to-end feature logic without modifying code is possible by applying read-only validation to trace flows from UI through service and persistence layers, checking for undocumented side effects and spec conformance.

Does this feature review process automatically fix the logic gaps it finds?

The feature review process does not automatically fix logic gaps; it enforces strict read-only operation to identify spec mismatches and emits structured findings for follow-up fix orchestration workflows.

What is the best way to check spec compliance for a new code change before merging?

Checking spec compliance before merging is best achieved through read-only feature auditing that discovers PR context and ADRs, traces end-to-end flows across layers, and outputs structured findings on validation and completeness.