What is a P1 severity finding in an agent-native review?

A P1 finding represents a core workflow that is completely blocked from automation, such as a critical feature that only exists as a UI component with no API backend, effectively locking out AI agents.

What does it mean for code to be 'agent-native'?

Agent-native code ensures that any action a human can take in a UI is also available to an AI agent via an API or tool, and any data a human sees is also queryable by an agent.

How does this skill identify violations?

It scans your code for UI elements like buttons or forms and checks if there are corresponding backend endpoints, mutation functions, or CLI tools that allow an agent to perform the same logic.

Can I use this for security auditing?

While its primary focus is on AI accessibility and parity, it helps identify hidden data endpoints or inconsistent access patterns that improve overall system architecture and transparency.

Agent-Native Reviewer

Name: Agent-Native Reviewer
Author: jovermier

byjovermier

API 개발

Audits codebase architecture to ensure all features and data accessible to humans are equally accessible to AI agents via APIs and tools.

소개

The Agent-Native Reviewer is a specialized auditing tool designed for the era of AI-integrated software development. It analyzes your code to identify 'AI accessibility' gaps, ensuring that every action a user can perform in a UI has a corresponding API, CLI, or tool equivalent. By enforcing the principle of agent parity, this skill helps developers build software that is inherently automation-friendly, identifying UI-only bottlenecks, hidden data silos, and incomplete workflows that prevent AI agents from successfully navigating and interacting with the system.

주요 기능

Automated Severity Grading: Categorizes findings from P1 (Critical) to P3 (Nice-to-Have) based on automation impact.
Workflow Completeness Analysis: Detects multi-step processes that require manual human-only UI interaction.
Data Visibility Scanning: Ensures all user-visible information is queryable via agent-accessible endpoints.
0 GitHub stars
Action Parity Auditing: Identifies user actions that lack REST, GraphQL, or CLI equivalents.
Event & Notification Checks: Verifies that agents can subscribe to the same real-time updates as users.

사용 사례

Standardizing developer practices to build 'agent-first' applications that support seamless AI integration.
Pre-release architectural audits to ensure high AI-readiness and automation support.
Identifying legacy UI-only features that need to be exposed via API for agent workflows.

소개

주요 기능

Automated Severity Grading: Categorizes findings from P1 (Critical) to P3 (Nice-to-Have) based on automation impact.
Workflow Completeness Analysis: Detects multi-step processes that require manual human-only UI interaction.
Data Visibility Scanning: Ensures all user-visible information is queryable via agent-accessible endpoints.
0 GitHub stars
Action Parity Auditing: Identifies user actions that lack REST, GraphQL, or CLI equivalents.
Event & Notification Checks: Verifies that agents can subscribe to the same real-time updates as users.

사용 사례

Standardizing developer practices to build 'agent-first' applications that support seamless AI integration.
Pre-release architectural audits to ensure high AI-readiness and automation support.
Identifying legacy UI-only features that need to be exposed via API for agent workflows.