deep-research

1# Deep Research
2 
3Create high-fidelity research reports with strict format control, evidence mapping, source governance, and multi-pass synthesis.
4 
5## Architecture: Lead Agent + Subagents
6 
7```
8Lead Agent (coordinator — minimizes raw search context)
9  |
10  P0: Environment + source policy setup
11  |
12  P1: Question and claim map (decision questions, evidence routes, stop rules)
13  |
14  Dispatch ──→ Subagent A ──→ writes task-a.md ──┐
15           ──→ Subagent B ──→ writes task-b.md ──┤ (parallel)
16           ──→ Subagent C ──→ writes task-c.md ──┘
17  |                                               |
18  |     research-notes/  <────────────────────────┘
19  |
20  P2: Dispatch + investigate → evidence packets
21  P3: Citation registry + source governance
22  P4: Evidence-mapped outline with counter-evidence and unknowns
23  P5: Draft from evidence packets; reopen decisive originals
24  P6: Counter-review (claims, confidence, alternatives)
25  P7: Verify → asset check and catalog registration
26```
27 
28**Context discipline:** Keep raw search-result noise in task workspaces. Pass evidence packets to the lead agent, including locators and short source excerpts. Notes are routing aids, not authorities: the lead agent must open the original source for every load-bearing claim, conflicting claim, and exact figure/date/quotation used in the report.
29 
30## Run and asset contract — every invocation
31…