AD cn-model-gateway
国产大模型统一 MCP 服务器，通过标准 JSON-RPC 2.0 协议为 Claude Code / Cursor / Cline / n8n 等 18+ Agent 框架提供 DeepSeek、通义千问、智谱 GLM、Kimi、腾讯混元、火山豆包、MiniMax、零一万物、百川智能、阶跃星辰十家模型的统一调用接口。新增 5 个非 MCP 框架适配器：LangChain Tool、AutoGPT Plugin、CrewAI Tool、Coze 插件、Dify 工具节点，实现从 MCP 生态到全 Agent 生态的扩展。内置模型性能基准测试套件（50 道题库、6 维度评分、雷达图对比、历史追踪）和 Token 价格实时追踪（价格抓取、变更通知、趋势图、成本预测）。支持 8 个 MCP 工具（ask_model/describe_image/embed_text/rerank/audio_transcribe/video_understand/list_providers/health_check）、资源读取（配置/使用统计）、预置 prompt 模板（代码审查/翻译），内置统一错误映射、流式 SSE 输出、使用量统计、硬件感知并发控制。auto 模式支持能力画像排序 + 自动故障转移（超时/失败切备用）；API key 支持环境变量优先读取；SQLite 启用 WAL 模式支持多 Agent 框架并发写入；支持多模态视觉模型（Qwen-VL/GLM-4V/豆包视觉）和图片理解（describe_image）；支持 Function Calling / Tool Use（ask_model 传入 tools 参数）；新增文本向量嵌入（embed_text）、文档重排序（rerank）、语音转文字（audio_transcribe）、视频理解（video_understand）四个新工具。config.json 填写 api_key 即可启动，无需 GPU、不做微调、不做私有部署，只做标准 MCP 协议网关。
国产大模型统一 MCP 服务器，通过标准 JSON-RPC 2.0 协议为 Claude Code / Cursor / Cline / n8n 等 18+ Agent 框架提供 DeepSeek、通义千问、智谱…
As a process D 46/100 · Unfinished process — weak spots: result and completion, when it triggers, inputs and preconditions
How to improve
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 1
✓ No critical or high findings
Medium and low: 1
-
low Secrets in code
secret-high-entropy-token_core/cache.py:38High-entropy token-like string (may be an id, hash or a credential) (quoted — discussed, not commanded)_embedding_model = SentenceTransformer('para…-v2')quoted
Files scanned: 59. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
✓ No remarks against the Agent Skills spec
Process rating: all ten parameters 46/100
- 0Result and completion. Does not say what the result is
- 0Inputs and preconditions. Does not say what the process needs to start
- 0Failures and branches. Linear process with no failure handling
- 0Progress reporting. Says nothing while it works
- 20When it triggers. No condition that starts the skill
- 60Tools and files. Uses tools (python) that frontmatter does not declare
- 100Steps. 24 steps
- 100Consistency. Name and required fields are in place
- 100Execution cost. Instruction body is 2682 tokens
- 100Running it twice. No mutating operations
- low 15 top-level sections: this looks like several domains in one skill
- low The skill ranks results itself: that belongs to the system behind the tool, not the model
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +5Description has no quoted example phrases that should trigger the skill
- +4Description does not say when NOT to use the skill (false activations)
- +3Description length 867: 120–800 characters recommended
- +3Output format is not stated: the model decides each time
- -217 emoji in the instructions: noise for the model
- +1No license
- +2Single-language instructions
- +4Structure: 22 headings
- +3Step-by-step instructions: 24 items
- +4Has examples (6 code blocks)
Quality base 70; lint remarks subtract, signals add up to 100. Result: 79.