skills.rest
tapway
tapwayOFFICIAL
Explore all skills by this author →
WebsiteGitHub

search-router

Classifies search queries by intent and routes them to Exa or last30days backends.

1|Updated Jun 21, 2026
Search Tags:#research#sentiment analysis#query classification#exa#intent detection#search routing#reddit search
One-click install
npx skills add https://github.com/tapway/shogun-os --skill search-router-tapway
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: search-router
Source: https://github.com/tapway/shogun-os/tree/main/skills/hermes/search-router
Command: npx skills add https://github.com/tapway/shogun-os --skill search-router-tapway

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Research queries often get sent to the wrong search backend, wasting time and returning irrelevant results — factual lookups drown in forum noise while opinion questions return marketing pages. This Skill classifies each query as FACTS, OPINION, or MIXED and routes it to the correct backend before any research begins. ## Core Features & Use Cases - Intent Classification: Categorizes queries as FACTS (scores, docs, release notes), OPINION (comparisons, sentiment, recommendations), or MIXED (broad research requests). - Backend Routing: Sends factual queries to Exa via web_search, opinion queries to the last30days script covering Reddit and Hacker News, and runs both in parallel for mixed queries. - Query Planning: Generates a --plan JSON with intent, freshness mode, and subqueries for named-entity searches in last30days. - Use Case: For the query "Cursor vs Copilot which is better", the Skill classifies it as OPINION and routes only to last30days, returning real Reddit and HN discussions instead of vendor blog posts. ## Quick Start Classify my research query and search the right backend for "Hermes Agent pricing and user reviews".

Frequently Asked Questions about search-router

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

FAQPage Schema
How do I route search queries between factual and opinion sources?▼

Classify the query intent first: factual queries (scores, docs, release notes) go to Exa via web_search, while opinion queries (comparisons, sentiment) go to last30days covering Reddit and Hacker News. Mixed queries run both backends in parallel.

When should I use Exa vs Reddit search for research?▼

Use Exa for objective, verifiable information like documentation, prices, and changelogs. Use last30days (Reddit, Hacker News) for subjective questions like tool comparisons, recommendations, and community sentiment about a product.

Can I run factual and sentiment searches in parallel?▼

Yes, MIXED queries trigger both backends simultaneously since they have no dependency. Exa returns in about 2 seconds and last30days in 5-35 seconds, with results presented side-by-side with source attribution.

What are the prerequisites for using query routing with last30days?▼

You need the web_search tool with an Exa backend, the last30days skill installed at /tmp/last30days-skill/, and Python 3.12 to run its script. Named-entity queries also benefit from generating a --plan JSON first.

When should I not use a search router?▼

Skip routing for trivial factual lookups like release dates, where a direct web_search suffices. Also avoid it for internal knowledge queries, which should go to gbrain-query instead of external search backends.

Related Skills

Explore More Education & Research Skills→
udacityudacity
28

TODO-skill-name

Classify support tickets by urgency and category for routing systems.

Official
Intermediate
goodnight000goodnight000

rag-implementation

Build RAG systems with vector stores, embeddings, retrieval, and reranking.

Community
Advanced
meiiiemeiiie
4

deep-research

Cross-reference factual claims against multiple authoritative sources with citations.

Community
Advanced
miptah21miptah21
2

search-strategy

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

Community
Advanced