agent-native-reviewer

Review code and designs for UI-agent action parity and tool primitives.

49|4|Updated Feb 12, 2026

Search Tags:#developer tools #code review #system design #parity #agent-native #agent accessibility

Ready to run

npx skills add https://github.com/gvkhosla/compound-engineering-pi --skill agent-native-reviewer-gvkhosla

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: agent-native-reviewer
Source: https://github.com/gvkhosla/compound-engineering-pi/tree/main/skills/agent-native-reviewer
Command: npx skills add https://github.com/gvkhosla/compound-engineering-pi --skill agent-native-reviewer-gvkhosla

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill ensures that any action a user can perform through the UI can also be performed by an AI agent, preventing features from being inaccessible to automated workflows.

Core Features & Use Cases

Action Parity Verification: Checks if UI actions have corresponding agent tools.
Context Parity Analysis: Ensures agents have access to the same data as users.
Tool Design Review: Validates that agent tools are primitives, not complex workflows.
Use Case: After adding a new "publish to feed" button in the UI, this Skill verifies that an agent can also trigger the "publish to feed" action using a dedicated tool.

Quick Start

Run the agent-native-reviewer to check the parity of the new user profile editing feature.

Frequently Asked Questions about agent-native-reviewer

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

FAQPage Schema

How do I check if my agent tools are designed as primitives and not complex workflows?▼

Context parity analysis ensures agents have access to the same data context as users. It reviews system prompts and data context to identify and rectify issues like context starvation, preventing automated workflows from missing critical information.

How do I check if my agent tools are designed as primitives and not complex workflows?▼

Tool design review validates that agent tools are primitives, not complex workflows. It analyzes tool definitions to ensure they provide fundamental capabilities that agents can compose, rather than rigid automated sequences.

Why does my agent fail to trigger actions that are available in the application UI?▼

Agent action failures often stem from orphan features where UI elements lack corresponding agent tools. Reviewing application designs for agent-native principles identifies these gaps and rectifies sandbox isolation issues to restore functional parity.

What is agent-native code review and when do I need it?▼

Agent-native code review is a process that enforces agent-native principles by analyzing UI elements, agent tools, system prompts, and data context. You need it when adding UI features to ensure automated workflows maintain equal capabilities with users.

Does agent-native review work without external dependencies?▼

Yes, agent-native review operates without external dependencies. It directly analyzes your application designs, shared workspace adherence, and tool definitions to validate parity between user interface actions and agent capabilities.