tech-diagnosis

Run technical diagnosis by collecting code and git measurements, internal documents, and GitHub data to generate a data-driven report on architecture health, engineering org operations, and technology strategy. It uses a verification loop to validate facts before finalizing the analysis.

Yes, you can scope technical diagnosis to specific modes including architecture, org, roadmap, delta, or metrics. The roadmap mode focuses on migration guidance and EOL detection by optionally researching web sources and GitHub issues for technology strategy insights.

The best way to verify codebase metrics is through the fact sheet verification loop, which presents measured codebase scale, quality risk signals, and security or build deployment metrics to the user for correction before proceeding to the integrated architecture, org, and strategy analysis.

No, GitHub and web research for EOL and migration guidance are optional inputs. The technical diagnosis primarily relies on direct code and git measurements alongside internal documents, strictly labeling code-validated claims versus document-only claims when handling missing or conflicting data sources.

Structure discovery metadata using YAML frontmatter to define the scope and parameters for the technical diagnosis. This metadata drives the collection of codebase metrics, internal documents, and optional external research, ensuring the verification loop targets the correct architecture or org dimensions.

When missing or conflicting data sources occur during technical diagnosis, the tool strictly labels claims as code-validated or document-only and handles gaps explicitly. It cannot fabricate missing metrics, relying instead on the user verification loop to correct assumptions before generating the final leadership report.