Mr-remon219/dsh-search-boost

Mr-remon219★ 8JavaScript最后同步: 2026-08-15

在 GitHub 打开

The plunge for dsh to boost model's search ability.

README 摘要

dsh-search-boost Search boost for DeepSeek Harness (DSH): multi-engine fused search, focused page fetching, X (Twitter) search, step-mode deep research, parallel multi-agent research, and an injected proactive-search policy. A bundle plugin for DSH that upgrades the built-in web search and registers a family of search tools: - Free-by-default engines run in parallel : Antigravity CLI / Bing / DuckDuckGo (all keyless), with keyed Tavily / Brave / Exa joining when keys are present. - Fused multi-engine ranking with cross-engine co-occurrence scoring and time-decay freshness. - Deep research driven by the main agent, and parallel research fanned out to native DSH subagents. Features Capability Description Built-in web search upgrade Registers a WebSearchProvider and patches searchProvider , so the built-in web search runs on this plugin's free-first engine chain (native citation cards preserved) fused search Multi-engine fused retrieval: free engines run in parallel (Antigravity CLI / Bing / DuckDuckGo — all keyless, two are plain curl scrapes), keyed engines join when keys exist (Tavily / Brave / Exa). Complexity routing, Grok-style query preprocessing ( site: / OR / quotes), hard do…

在 GitHub 查看完整 README →
搜索deepseek-harnessdsh-pluginweb-searchagentsearch

分类