hybrid-deep-search
от scsun1978
Search & ResearchКратко
Hybrid Deep Search - Intelligent routing between Brave API (free, fast) and OpenAI Codex (deep analysis, paid).
Описание
Что делает этот навык
Routes search queries between Brave API and OpenAI Codex based on query complexity. Simple factual questions go to Brave (free, sub-2-second responses), while comparison, analysis, and multi-source questions route to OpenAI Codex. Mode selection is automatic but can be overridden with --mode flags.
Auto-routing between a free fast API and a paid reasoning model keeps costs proportional to query complexity without requiring manual tool selection.
Когда его использовать
- Looking up quick facts without spending API credits
- Comparing two software frameworks with in-depth analysis
- Searching for recent academic papers on a research topic
- Running batch searches across mixed-complexity queries
- Pulling latest news with automatic backend selection
Пример сценария
User asks: compare LangChain vs LlamaIndex in detail
- 1
Analyze query for complexity keywords such as 'compare' and 'in detail'
- 2
Score query above threshold (3+) using NLP keyword and length rules
- 3
Route to OpenAI Codex (gpt-5-codex) for deep analysis
- 4
Fetch multi-source synthesis via OpenAI Chat Completions API
- 5
Return structured markdown comparison
Detailed multi-source comparison of LangChain and LlamaIndex covering architecture, use cases, and tradeoffs
Требования
- OPENAI_API_KEY environment variable
- OpenAI account at platform.openai.com
Установка
clawhub install scsun1978/hybrid-deep-searchВыполните команду CLI или вставьте промпт в Claude Code, Cursor, Codex — любой кодинг-агент, которым пользуетесь.