skills.rest
miptah21
miptah21COMMUNITY
Explore all skills by this author →
GitHub

search-strategy

Convert natural-language questions into parallel source-specific searches and synthesize ranked results.

2|Updated Apr 8, 2026
Search Tags:#semantic search#information retrieval#knowledge base#project tracking#deduplication#query decomposition#result ranking
One-click install
npx skills add https://github.com/miptah21/skills --skill search-strategy-miptah21
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: search-strategy
Source: https://github.com/miptah21/skills/tree/main/.agents/skills/search-strategy
Command: npx skills add https://github.com/miptah21/skills --skill search-strategy-miptah21

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill solves the problem of producing fast, accurate answers by decomposing a natural-language question into targeted searches across multiple information sources and then synthesizing ranked results.

Core Features & Use Cases

  • Query decomposition by question type: classifies the query (decision, status, document, person, factual, temporal, exploratory) to choose the most effective search priorities.
  • Parallel, source-specific search orchestration: generates semantic or keyword queries per source and maps enterprise-style filters into source syntax.
  • Relevance ranking with authority and freshness: scores results using weighted factors and applies an authority hierarchy to prefer the most trustworthy evidence for the question type.
  • Ambiguity handling and fallbacks: asks a focused clarifying question only when interpretations would materially change results, otherwise broadens queries or skips unavailable sources with graceful degradation.

Quick Start

Use search-strategy to answer: "What did we decide about the API migration timeline?" by running parallel chat, wiki, and project-tracker searches, then merging and deduplicating ranked results into one coherent response.

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 project trackers with one natural-language query?▼

Query decomposition splits a natural-language question into targeted sub-queries, classifying it by type (decision, status, document, person, factual, temporal, exploratory) to determine search priorities. This ensures the most effective search strategy is applied for each specific question.

How are search results deduplicated and ranked when pulling from heterogeneous information sources?▼

Relevance ranking scores results using weighted factors like authority and freshness, applying an authority hierarchy to prefer the most trustworthy evidence for the question type. Deduplication synthesizes overlapping results across heterogeneous systems into one coherent response.

Does the search strategy use semantic search or keyword search for information retrieval?▼

The search strategy uses both semantic and keyword searches. It selects between semantic-vs-keyword queries based on the specific source being searched and translates enterprise-style filters into the appropriate source syntax for accurate information retrieval.

What happens when search queries are ambiguous or specific information sources are unavailable?▼

Ambiguity handling asks a focused clarifying question only when interpretations would materially change results. When sources fail, the system applies fallback query broadening and skips unavailable sources with graceful degradation to maintain functionality.

Can I use this for exploratory research or is it limited to specific factual questions?▼

You can use this for exploratory research as well as decision, status, document, person, policy, factual, and temporal queries. The query decomposition classifies the question type to choose the most effective search priorities for any supported query.

Related Skills

Explore More Data & Analytics Skills→
openyakopenyak
704

search-strategy

Decompose natural language questions into parallel source-specific searches across connected data stores.

Official
Advanced
OpenLAIROpenLAIR
1.0k

dataset-discovery

Find and rank datasets from HuggingFace, OpenML, GitHub, and papers.

Official
Intermediate
ImBIOSImBIOS
1

learning-systems

Convert task outcomes into learning signals for swarm decomposition and pattern tracking.

Community
Advanced
msd-dev-labmsd-dev-lab

agent-memory

Store and retrieve user memories across conversations with categorized folders.

Official
Basic