search

Search connected MCP sources with parallel queries and synthesized results.

1|Updated Mar 19, 2026
One-click install
npx skills add https://github.com/cy-wali/knowledge --skill search-cy-wali
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: search
Source: https://github.com/cy-wali/knowledge/tree/main/enterprise-search/skills/search
Command: npx skills add https://github.com/cy-wali/knowledge --skill search-cy-wali

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Unified search across chat, email, cloud storage, project trackers, CRM, and knowledge bases by decomposing queries, running parallel searches, and synthesizing results.

Core Features & Use Cases

  • Determine available sources and identify connected MCP tools from the tool list.
  • Parse user queries to extract intent, entities, time constraints, and source hints, then decompose into targeted sub-queries.
  • Execute sub-queries in parallel across all connected sources and rank/deduplicate results to present a single synthesized answer.
  • Support scenarios like locating decisions, documents, conversations, or status updates across multiple tools.

Quick Start

Ask it to find a document or answer across all connected sources with a single query.

Frequently Asked Questions about search

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

FAQPage Schema
How do I search across multiple connected sources like email and cloud storage at the same time?

Unified search across connected sources is achieved by decomposing your query into targeted sub-queries, running parallel searches across chat, email, cloud storage, project trackers, CRM, and knowledge bases, and synthesizing ranked, deduplicated results.

Can I locate a specific decision or document across different project trackers and chat tools?

Yes, you can locate decisions, documents, or conversations across project trackers and chat tools by parsing your query to extract intent, entities, and time constraints, then executing parallel searches across all connected MCP sources.

What is the best way to find information across connected knowledge bases without checking each tool individually?

The best way to find information across knowledge bases is to use a single query that automatically identifies available MCP sources, decomposes the search intent, and returns a single synthesized answer with ranked and deduplicated results.

Does this parallel search approach work with CRM and cloud storage platforms simultaneously?

Yes, parallel search works with CRM and cloud storage platforms simultaneously by identifying connected MCP tools from your tool list and executing decomposed sub-queries across all available sources to synthesize a unified result.

How does query decomposition work when retrieving information across multiple data sources?

Query decomposition parses your input to extract intent, entities, time constraints, and source hints, breaking it into targeted sub-queries that run in parallel across connected sources before ranking and deduplicating the retrieved information.

What are the limitations of searching across all connected MCP sources in one query?

The search is limited to connected MCP sources available in your environment. It can only identify and query tools currently connected, synthesizing results from available chat, email, cloud storage, project trackers, CRM, and knowledge base integrations.