Research Planner
The orchestration brain that decides agent sequence, checkpoints, and stop conditions.
The research planner agent is the control layer for complex research runs.
What it does
It decides which agents to run, in what order, and under what stop criteria. It also sets verification checkpoints and loop policies to avoid static one-pass workflows.
Output
A structured orchestration plan with:
- objective and key questions
- agent sequence and handoffs
- stop conditions (confidence, contradiction thresholds)
- verification checkpoints
Used by
Most useful in /plan-research, /deepresearch, and /iterate for non-trivial research tasks.