investigation-mode

Stalled progress from repeated build errors triggers an automatic pause to implementation, initiating a root cause analysis process that requires deep research and systematic testing before allowing coding to resume with a verified plan.

Root cause analysis for debugging consecutive failures is a systematic process that mandates deep research and isolation of issues, ensuring you understand the underlying error before resuming implementation rather than applying temporary patches.

You need to pause implementation for error handling and investigation when stall detection identifies two or more consecutive failures, automatically halting development to prevent compounding errors and enforce a structured debugging workflow.

To resume coding after root cause analysis, you must present a verified plan and obtain explicit user approval for any significant changes in direction or the adoption of workarounds before implementation can safely continue.

Investigation mode requires explicit user approval for adopting workarounds or making significant changes in direction, ensuring controlled resumption of development only after a verified plan is established and agreed upon.

Automated stall detection is limited to triggering after two or more consecutive failures, meaning isolated or intermittent errors may not initiate the root cause analysis process, requiring manual intervention to pause and investigate progress.