deep-debugger

Root-cause analysis on multi-file codebases requires a phase-driven approach: read existing evidence first, build an evidence map across modules, and produce a precise diagnosis before proposing fixes. This ensures traceable debugging across distributed components.

Evidence-based debugging is a structured investigation method that maps symptoms to root-causes using traceable proof rather than assumptions. You need it for integration failures where symptoms are distributed across multiple interacting components or pipeline stages.

Debug pipeline failures by following a structured diagnostic process: gather logs and traces, construct an evidence map of interacting components, and propose minimal remediation steps with explicit file paths and line references for safe resolution.

Traceability in distributed system diagnosis requires generating outputs with explicit file paths and line references throughout the investigation. This approach supports debugging reviews by mapping each piece of evidence back to specific locations in the codebase.

Yes, it works across multi-file codebases and pipelines specifically designed for deep-dives when logs, traces, and prompts indicate multiple interacting components. It handles integration failures where symptoms are distributed across modules.

Phase-driven evidence mapping may be excessive for single-file bugs with obvious causes where immediate code changes suffice. It is designed for complex, multi-file codebases and pipeline failures requiring thorough root-cause investigation before remediation.