skills.rest
sun2443
sun2443COMMUNITY
Explore all skills by this author →
GitHub

search-strategy

Decompose natural-language questions into source-specific searches with relevance ranking and deduplication.

Updated Apr 8, 2026
Search Tags:#semantic search#deduplication#enterprise search#query decomposition#fallback strategy#relevance ranking#multi-source retrieval
One-click install
npx skills add https://github.com/sun2443/designer-skills --skill search-strategy-sun2443
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: search-strategy
Source: https://github.com/sun2443/designer-skills/tree/main/enterprise-search/skills/search-strategy
Command: npx skills add https://github.com/sun2443/designer-skills --skill search-strategy-sun2443

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill solves the problem of finding the right information quickly across multiple enterprise systems when a user only has a natural-language question.

Core Features & Use Cases

  • Query decomposition: Breaks a single question into query components (keywords, entities, intent signals, constraints, and negations) based on query type such as decision, status, document, person, policy, temporal, or exploratory.
  • Source-specific translation: Generates semantic and keyword variants per connected source and maps enterprise-style filters into each source’s syntax.
  • Ranking, deduplication, and fallback: Scores relevance using weighted factors (keyword match, freshness, authority, completeness), deduplicates overlapping results, and broadens or shifts strategy when sources fail or return nothing.
  • Use case: When asked, "What did we decide about the API migration timeline?", it searches chat/email, knowledge base/wiki, and the project tracker in parallel, then synthesizes a single coherent answer from the highest-signal results.

Quick Start

Ask the assistant: "Search across all connected sources and summarize what we decided about the API migration timeline."

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 enterprise systems using a natural language question?▼

Multi-source enterprise search decomposes natural-language questions into targeted, source-specific queries, retrieving and synthesizing ranked, deduplicated answers from connected systems like chat, knowledge bases, and project trackers.

How does query decomposition handle different types of enterprise searches?▼

Query decomposition breaks questions into keywords, entities, intent signals, constraints, and negations based on query type, enabling precise retrieval for decision, status, document, person, factual, temporal, or exploratory searches.

What is the best way to rank and deduplicate search results from different sources?▼

Ranking and deduplication score relevance using weighted factors like keyword match, freshness, authority, and completeness, merging overlapping results from multi-source retrieval into a single coherent answer.

What happens when an enterprise search source is unavailable or returns no results?▼

When sources fail or return empty, a fallback strategy broadens the query or shifts retrieval tactics, applying alternative semantic or keyword variants to ensure multi-source search continuity.

Can I map enterprise-style filters to each connected source's specific query syntax?▼

Yes, source-specific translation generates parallel semantic and keyword variants while mapping enterprise-style filters directly into each source's native query syntax for accurate multi-source retrieval.

How do I synthesize a coherent answer from parallel search results across chat, wiki, and project trackers?▼

Synthesizing parallel retrieval results involves scoring relevance by weighted factors, deduplicating overlapping items, and extracting the highest-signal documents to formulate a coherent answer.

Related Skills

Explore More Product & Management Skills→
kousenkousen
10

osquery

Translate natural language questions into Osquery SQL queries and parse JSON results.

Community
Advanced
anthropicsanthropics
23.4k

knowledge-synthesis

Combine search results from multiple enterprise sources into deduplicated answers with source attribution.

Official
Advanced
giosuetedeschi-specgiosuetedeschi-spec

rag-engineer

Design and optimize Retrieval-Augmented Generation pipelines for improved context retrieval.

Community
Advanced
zzafergokzzafergok
1

search-strategy

Decompose user questions into parallel, source-specific search queries.

Community
Advanced