Analyze dlab session runs

To analyze dlab session logs, open the session directory and examine main.log and _opencode_logs to trace orchestrator behavior, verify parallel agent results, and identify execution issues across runs.

Auditing parallel agent results involves inspecting the session's state.json and parallel run summary files to evaluate performance, diagnoses, and how well prompts and domain skills aligned with outcomes.

Yes, you can review decision-pack performance by navigating completed session directories to compare outputs, assess uncertainty handling, and conduct capacity planning based on orchestrator and parallel run summaries.

Check uncertainty handling and diagnoses by inspecting the _opencode_logs and summary files within the session directory to evaluate how parallel agents processed diagnoses and managed uncertain outcomes.

You need to inspect the session directory's state.json, main.log, and _opencode_logs to verify orchestrator behavior, trace execution flow, and validate the alignment of prompts with results.