dk-local-inspect

Run local diagnostic scripts from the repository root's taskit/taskit-backend directory using testing_tools. This executes local diagnostics to inspect tasks, boards, specs, or reflections and captures their current state in a structured format for debugging.

Supported diagnostic targets include task, spec, board, reflection, and snapshot. Diagnosing these targets allows you to investigate failures, validate board state, or audit reflections before making project decisions.

Yes, you can use the json output mode to get structured diagnostic results for LLMs. The local diagnostics also support brief and full output modes to fit different debugging scenarios and verbosity needs.

Inspecting specs requires no external dependencies, but you must execute scripts from a consistent repo root within the taskit/taskit-backend directory. This ensures the diagnostic scripts capture outputs predictably.

The best way to reduce guesswork is to run local diagnostic scripts to inspect task state directly. This provides predictable, structured outputs that help investigate failures and validate state before making decisions.