gsd-debug

Debugging complex issues across sessions requires persistent state management to save progress. This approach coordinates symptom gathering and checkpoint handling, allowing you to pause and resume deep investigations without losing context.

Subagent orchestration spawns specialized agents to investigate specific symptoms while isolating their context. This keeps the main assistant lean and protects user interaction from context bloat during complex debugging tasks.

Context isolation is maintained by spawning subagents for specific investigations and enforcing clear prompts. This prevents the main context from expanding excessively while preserving the state needed for continuation workflows.

Checkpoint handling allows you to save the state of a debugging investigation and resume it later. You can start a new investigation by describing the issue and desired outcome, then use continuation workflows to pick up where you left off.

The best way to coordinate symptom gathering is using an orchestration workflow that manages state and spawns subagents. This provides interactive prompting and deterministic task execution to guide you through complex debugging scenarios.

Debugging context overflows when all investigation data is kept in the main assistant. Using subagent isolation and context management prevents this, keeping user interactions lean while preserving necessary debugging state.