debug-task

Agent tasks fail to load skills or start in the wrong working directory due to infrastructure issues like catalog path bugs or workspace isolation gaps. Analyzing backend logs and task transcripts traces these detours to their specific root cause in the platform code or configuration.

Debug AI agent execution failures by reading actual agent debug logs and SQLite task databases before forming conclusions. This log-driven debugging approach prevents assumptions from file existence alone and traces failures to specific code or configuration root causes.

Yes, you can diagnose agent infrastructure issues across WSL and native Windows deployments. Cross-platform topology detection handles environments with separate data directories, preventing stale database reads when native Windows servers and WSL operate simultaneously.

The best way to find the root cause of browser startup failures is analyzing CDP proxy problems and agent wrapper scripts. Tracing the execution infrastructure reveals whether the failure stems from configuration gaps or browser control protocol issues.

Investigate missing files during agent task execution by analyzing backend logs and agent transcripts rather than relying on task messages alone. This identifies whether the issue stems from catalog path bugs or workspace isolation gaps in the infrastructure.

Cross-platform deployment limitations include stale database reads when WSL and native Windows environments maintain separate data directories. Analyzing the SQLite task database and wrapper scripts ensures accurate topology detection across isolated store agents.