Add summarization flag to quick_search: search-grounded quick summaries (Vertex AI-style) #1603
Description
Is your feature request related to a problem? Please describe.
No.
Describe the solution you'd like
Add an opt-in aggregated-summary mode to quick_search so clients can request a fast, search-grounded aggregated summary across top search results without running the full deep_research / write_report pipeline. This mirrors the Vertex AI “search grounding” pattern but returns only a single synthesized, citation-aware summary of the selected sources (no per-source summaries).
Motivation:
- Provides a fast, lower-cost option for assistants that need a single, grounded synthesis of recent web results.
- Reuses existing quick_search and writer logic with minimal API surface change and backwards compatibility.
- Useful for MCP integrations and interactive assistants that want a single summary instead of raw snippets or a full research report.