gpd-debug

Debug complex physics calculation discrepancies by orchestrating parallel subagents that isolate root causes through structured symptom input and evidence-driven diagnosis. This approach coordinates multiple investigations to identify conflicting hypotheses in iterative research workflows.

Systematic root-cause analysis for divergent numerical pipelines involves coordinating parallel subagents that carry state across context resets, enabling structured diagnosis where calculations diverge across simulations, analyses, and numerical pipelines with traceable debug sessions.

Start a physics debugging session by describing a new scenario with symptoms and instructing the orchestrator to spawn investigators. The system maintains state between sessions and across restarts, saving debug sessions locally under the .gpd/debug directory.

Yes, state persistence maintains context between sessions and across restarts for iterative research workflows. Debug sessions are saved locally, allowing orchestrated diagnostics to carry state across context resets and resume systematic root-cause analysis.

Run parallel root-cause discovery by orchestrating gpd-debugger agents that execute multiple investigations simultaneously to identify conflicting hypotheses. This coordinated approach handles symptom propagation and checkpoint handling while isolating root causes through evidence-driven diagnosis.

Yes, physics debugging operates without external dependencies, using modular agents and traceable debug sessions stored under .gpd/debug. The system provides documented root-cause reporting through evidence-driven diagnosis with structured symptom input and checkpoint handling.