360 Secure SkillHub
360 安全技能中心

LLM Router

该 Skill 提供 llm_router_client 命令行脚本,通过 AIsa 路由服务完成模型列表查询、单模型聊天、多模型对比与图像理解调用,并给出 AIsa 提供商配置指引;需配置 AISA_API_KEY,依赖 python3,随包附有 README 与脚本。

baofeng-tech
baofeng-tech
热度 43.55
最近更新 2026-05-11 17:45:14
分类:开发依赖:代码执行 / 第三方 API / 内部调用大模型应用场景:API 集成 / 工作流

安装技能

复制下面的话术发给你的 Agent,它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill: npx secure-skills-cli add llm-router
Metadata
name
llm-router
description
Use AIsa for model routing, provider setup, and Chinese LLM access. Use when: the user needs model configuration, provider guidance, or routing workflows. Supports setup and model operations.
author
AIsa
version
1.0.0
license
Apache-2.0
user-invocable
true
primaryEnv
AISA_API_KEY
requires
{"bins":["python3"],"env":["AISA_API_KEY"]}
metadata
{"aisa":{"emoji":"🤖","requires":{"bins":["python3"],"env":["AISA_API_KEY"]},"primaryEnv":"AISA_API_KEY","compatibility":["openclaw","claude-code","hermes"]},"openclaw":{"emoji":"🤖","requires":{"bins":["python3"],"env":["AISA_API_KEY"]},"primaryEnv":"AISA_API_KEY"}}

LLM Router

Use AIsa for model routing, provider setup, and Chinese LLM access. Use when: the user needs model configuration, provider guidance, or routing workflows. Supports setup and model operations.

When to use

  • The user needs model routing, provider setup, or Chinese LLM access.
  • The user wants one place for provider configuration or model selection.
  • The user wants setup guidance for AIsa-hosted model workflows.

High-Intent Workflows

  • Configure an AIsa provider path.
  • Inspect supported models or routing options.
  • Prepare a runtime for Chinese-model access.

Quick Reference

  • python3 scripts/llm_router_client.py --help

Setup

  • AISA_API_KEY is required for AIsa-backed API access.
  • Use repo-relative scripts/ paths from the shipped package.
  • Prefer explicit CLI auth flags when a script exposes them.

Example Requests

  • Help me configure AIsa for Qwen
  • List the supported routed models
  • Choose a model for Chinese long-form analysis

Guardrails

  • Do not ask for extra credentials beyond the shipped flow.
  • Do not advertise setup paths that the public bundle does not ship.
  • Keep setup instructions aligned with the actual runtime.