search-strategy

Decompose user queries into sub-queries for parallel searching across multiple sources.

4|4|Updated Dec 15, 2024

Search Tags:#synthesis #ranking #semantic-search #knowledge-base #multi-source-search #split-queries

Ready to run

npx skills add https://github.com/adrianliechti/wingman-chat --skill search-strategy-adrianliechti

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: search-strategy
Source: https://github.com/adrianliechti/wingman-chat/tree/main/skills/knowledge/search-strategy
Command: npx skills add https://github.com/adrianliechti/wingman-chat --skill search-strategy-adrianliechti

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Transforms a single natural language question into parallel, source-specific searches and synthesizes results from multiple connected sources.

Core Features & Use Cases

Query decomposition: Breaks a question into focused sub-queries per source (chat, knowledge base, wiki, task tracker).
Source translation & ranking: Translates queries into source-specific syntax, ranks results by relevance, and deduplicates results.
Ambiguity handling & fallbacks: Applies fallback strategies for missing sources and asks clarifying questions when needed.
Use Case: A product manager asks "What are the latest decisions on API changes?" and receives a ranked overview from chat threads, docs, and issue trackers.

Quick Start

Provide a natural language question and the system will decompose it into targeted searches across connected sources.

Frequently Asked Questions about search-strategy

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I search across multiple knowledge bases and chat threads at the same time?▼

Multi-source search solves this by decomposing a natural language query into source-specific sub-queries, distributing them in parallel across wikis, knowledge bases, and chat threads, then ranking and deduplicating results into one synthesized answer.

What is the best way to split a complex question into targeted sub-queries for different sources?▼

Query decomposition breaks down a complex question into focused sub-queries tailored per source. It translates queries into source-specific syntax for parallel execution across chat, task trackers, and wikis, applying fallback strategies for missing sources.

How does semantic search handle precise terms versus conceptual queries across a wiki?▼

Semantic search handles conceptual queries while keyword searches target precise terms. The system applies both methods during multi-source parallel execution, ranking and deduplicating the combined results to synthesize a single answer.

Can I use multi-source search if one of my connected sources like a task tracker is offline?▼

Yes, multi-source search applies fallback strategies for missing sources. It distributes parallel searches across connected environments and asks clarifying questions when needed, ensuring synthesis continues despite a source being unavailable.

When do I need query decomposition for knowledge base retrieval?▼

You need query decomposition when a single natural language question requires information scattered across multiple sources. It translates questions into source-specific syntax to extract ranked, deduplicated results from chat, wikis, and task trackers.