ce-agent-native-audit

Auditing agent-native architecture involves evaluating your codebase and system design against principles like Action Parity and Tools as Primitives. This skill uses parallel sub-agents to review the codebase and generates a scored report highlighting gaps and recommendations.

An agent-native architecture audit is a structured review evaluating your system's adherence to design principles like Action Parity and Tools as Primitives. It generates scored reports to identify architectural gaps and provide targeted improvement recommendations.

Yes, you can audit a specific principle. The skill supports custom principle audits, allowing you to target individual concepts like Action Parity for focused analysis and scoring instead of running a full comprehensive system review.

Parallel sub-agent code review works by concurrently analyzing different architectural principles. Each sub-agent evaluates a specific principle like Tools as Primitives against your codebase, enabling faster comprehensive analysis and generating detailed scores for each area.

No, the architecture audit requires no external dependencies. The skill operates independently using its internal scripts and references to analyze your codebase and generate the principle-based scored reports without additional packages.

The best way to evaluate agent-native principles is through a structured, principle-based review using parallel sub-agents. This approach concurrently scores adherence to concepts like Action Parity, providing a detailed report of gaps and actionable recommendations.