search-core

Orchestrate Gemini delegation, source trust evaluation, normalization, and deduplication for web searches.

1|Updated Nov 14, 2025
One-click install
npx skills add https://github.com/inchan/cc-skills --skill search-core
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: search-core
Source: https://github.com/inchan/cc-skills/tree/main/plugins/search/skills/search-core
Command: npx skills add https://github.com/inchan/cc-skills --skill search-core

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Search Core provides a reusable, centralized web search pipeline that 3 agents share, plus credibility filtering and deduplication to deliver trusted results quickly.

Core Features & Use Cases

  • Unified search flow: Gemini delegation via MCP or Bash, plus a self-contained fallback.
  • 3 search agents: official-docs, comprehensive, best-practice.
  • Source credibility: Tier 1-4 scoring system.
  • Result normalization & deduplication: Consistent structure, unique URLs.

Quick Start

Use search-core as the backbone for any search Skill; provide a query and optional tier to filter sources.

Frequently Asked Questions about search-core

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

FAQPage Schema
How do I deduplicate and normalize web search results?

Search Core provides unified web search logic that automatically deduplicates results by URL and normalizes output to a consistent format. It handles Gemini delegation, source credibility filtering across Tier 1-4 sources, and produces structured results ready for integration.

Can I evaluate source credibility when aggregating search results?

Yes. Search Core implements a Tier 1-4 source assessment system that scores result credibility. You provide a query and optional tier threshold to filter results by trustworthiness before returning normalized output.

How do I integrate Gemini with web search in a reusable way?

Search Core orchestrates Gemini delegation through MCP or Bash with a WebSearch fallback, eliminating the need to rebuild search logic per agent. Three search agents (official-docs, comprehensive, best-practice) share this backbone to handle prioritized search methods consistently.

What's the best way to handle multiple search strategies in a single pipeline?

Search Core manages three distinct search agents with different prioritized methods—Gemini via MCP first, Gemini via Bash second, WebSearch as fallback. Results merge through deduplication and normalization into a single deduplicated output, ensuring consistent formatting across strategies.

Do I need to normalize URLs and results manually for search integration?

No. Search Core handles URL normalization and result normalization automatically as part of its unified pipeline. It produces a defined TypeScript/JavaScript schema output, eliminating manual formatting work across downstream integrations.

What dependencies or setup does web search core require?

Search Core has no external dependencies and integrates via a defined TypeScript/JavaScript input/output schema. Provide a query string and optional tier filter; it delegates to available methods (Gemini via MCP, Bash, or WebSearch) and returns structured results.