【帖子标题】：I built a local-first hybrid router for AI Agent Skills (sub-20ms, zero tokens, runs on CPU)
【翻译标题】：我开发了一款面向AI智能体技能的本地区优先混合路由系统（20毫秒内响应·零token消耗·CPU运行）
【帖子正文】：
Hey everyone,
【翻译正文】：大家好，
If you use agentic workflows with custom skills or rules (Cursor rules, Claude Code slash commands, OpenCode, etc.), you have probably run into the routing trade-off:
【翻译】：如果你使用带有自定义技能或规则的智能体工作流（如Cursor规则、Claude Code斜杠命令、OpenCode等），很可能遇到过这种路由权衡：
Stuff every skill definition into the system prompt (destroys your context窗口 and degrades instruction-following).
【翻译】：要么把所有技能定义塞进系统提示（会挤爆上下文窗口并降低指令跟随效果）
Use an LLM router turn to classify the user prompt (costs money, wastes 1,000+ tokens, and adds 2+ seconds of network latency).
【翻译】：要么用LLM路由轮次分类用户提示（耗费资金、浪费1000+token、增加2秒以上网络延迟）
To solve this, I built Routed; an open-source, local-first hybrid router for agent skills that runs 100% offline on your CPU.
【翻译】：为此我开发了Routed——一个开源的、本地区优先的智能体技能混合路由系统，100%离线运行于你的CPU上。
GitHub: https://github.com/bshea-1/Routed
License: MIT
【翻译】：GitHub地址：https://github.com/bshea-1/Routed
许可证：MIT
How it Works Under The Hood
【翻译】底层原理
Routed indexes your installed skill directories and evaluates prompts through a 4-part hybrid scoring pipeline:
【翻译】：Routed会索引已安装的技能目录，并通过4部分混合评分管道评估提示：
-
Dense Vector Embeddings (60%): Runs quantized ONNX models (Arctic Embed S / MiniLM) locally on CPU.
【翻译】：* 稠密向量嵌入（60%）：在CPU上本地运行量化ONNX模型（Arctic Embed S / MiniLM） -
Lexical BM25 (25%): Okapi BM25 for strict keyword relevance.
【翻译】：* 词法BM25（25%）：采用Okapi BM25算法实现严格关键词关联 -
Exact / Alias Match (10%): Direct command and alias matching.
【翻译】：* 精确/别名匹配（10%）：直接命令与别名匹配 -
Metadata (5%): Recency and usage heuristics.
【翻译】：* 元数据（5%）：基于时效性和使用频率的启发式规则
The entire lookup completes in under 20ms without sending a single byte of prompt data over the wire.
【翻译】：整个查找过程在20毫秒内完成，且不会通过网络传输任何提示数据。
Supported Environments
【翻译】支持环境
Routed auto-detects and injects adapters into:
【翻译】：Routed自动检测并注入适配器至以下平台：
Cursor, Claude Code, LM Studio, Ollama, Antigravity IDE, Windsurf, OpenCode, Continue, Codex, and I just dropped support for MCP Servers!!
【翻译】：Cursor、Claude Code、LM Studio、Ollama、Antigravity IDE、Windsurf、OpenCode、Continue、Codex，而且刚刚新增了对MCP服务器的支持！！
And although v1.0 dropped last night, I just shipped v1.1.0 with two major additions based on early feedback:
【翻译】：虽然v1.0昨晚才发布，但根据早期反馈我又推出了v1.1.0，包含两项重大更新：
Model Context Protocol (MCP) Server (routed mcp):
【翻译】模型上下文协议(MCP)服务器（routed mcp）：
Instead of loading 20+ tool schemas into your GPU’s context窗口, your local model only sees a single route_skill tool. Routed executes on CPU, selects the exact skill needed, and injects only that schema on demand.
【翻译】：无需向GPU上下文窗口加载20+工具架构，本地模型只需处理单个route_skill工具。Routed在CPU执行，精准选择所需技能并按需注入对应架构。
Native Multilingual Understanding:
【翻译】原生多语言理解：
The embedding pipeline now natively understands input across 100+ languages (German, Spanish, French, Japanese, etc.) and automatically decomposes compound nouns (like German Speicherleck), mapping prompts directly to the correct skill without language tags or manual translation.
【翻译】：嵌入管道现原生支持100+种语言输入（德语、西班牙语、法语、日语等），并能自动分解复合名词（如德语Speicherleck），无需语言标签或人工翻译即可将提示直接映射到正确技能。
Usage
【翻译】使用方式
Inside your agent chat, you can just use /route to trigger the best skill(s) dynamically. It also handles compound intents (e.g., matching multiple skills when a prompt asks for two distinct tasks).
【翻译】：在智能体聊天界面中，只需输入/route即可动态触发最佳技能。还能处理复合意图（例如当提示要求两项不同任务时匹配多个技能）。
Pre-built binaries are available on GitHub Releases (macOS .pkg, Linux .deb, Windows .exe), or you can build it from source via Node.
【翻译】：GitHub Releases提供预编译二进制文件（macOS .pkg、Linux .deb、Windows .exe），也可通过Node从源码构建。
Check it out and let me know what you think or if there are other environments you would like added!
【翻译】：欢迎试用并告诉我你的想法，或者提出需要增加的其他环境支持！