Register a single follow-up task discovered mid-investigation. Use add_tasks_bulk() instead if defining the full plan upfront.

FIRST CALL: Register your entire investigation plan as a DAG in one atomic call. All tasks are validated before any are inserted. Returns a prioritised execution plan showing which tasks to start with.

Resolve a task after completing work on it. Always provide a finding for completed/skipped/blocked. Completing a task automatically unblocks any dependents. Can also rewire dependencies (pending tasks only) or update a finding without changing status.

CALL AFTER EVERY update_task: Returns all tasks whose dependencies are fully resolved, sorted by priority (high → medium → low). These are the tasks you should work on next. Empty list means either all done or everything is blocked.

Get a full snapshot of the entire DAG — all tasks, their current status, findings, dependencies, and overall progress. Use when you need to review the full picture or check what is blocked.

FINAL CALL: Close out the investigation. Blocks and returns actionable instructions if any tasks are still pending or in_progress. Returns a full grouped summary (by status and category) once everything is resolved. Call this when you believe all tasks are done.

Reopen a blocked task after its blocker has been resolved. Resets the task to pending, preserves the original finding in history, and recalculates its dependency state. Provide a reason explaining what changed.

Clear all tasks and start a new investigation session. Irreversible — all tasks and findings are permanently deleted. Only use when starting a completely new investigation.