triage-architecture

Automating an architecture audit involves detecting your ticketing system, caching tickets to disk, and spawning parallel agents to assess safety, correctness, maintainability, and completeness, producing per-cluster tickets and a project map for remediation.

Multi-cluster codebase auditing is a process that orchestrates focused sub-agents to scrutinize safety, correctness, maintainability, and completeness, surfacing root architectural concerns and generating cross-cluster notes to weave insights back into existing tickets.

To run an architecture audit on a ticketed project, initiate the skill in create mode; it automatically detects the ticket system, caches tickets offline, and coordinates parallel risk-analysis agents to generate consolidated findings and a project map.

Yes, you can use parallel agents for code audits to assess maintainability and correctness simultaneously, as the workflow spawns four focused sub-agents dedicated to safety, correctness, maintainability, and completeness to scrutinize the codebase in parallel.

No specific ticketing system is required, but the architecture auditing workflow applies to projects with a formal ticketing system; it automatically detects the project's ticket system and caches tickets to disk for offline analysis during the audit.

A key limitation when auditing architecture across multiple code clusters is that the workflow requires a formal ticketing system to function, as it relies on detecting and caching tickets to disk to produce per-cluster tickets and cross-cluster notes for remediation.