crc-lnm-medical-agent
# CRC-LNM Medical Agent 1.0.12
这是面向 ModelScope 托管 STDIO 的完整懒加载纯 NumPy 单模型版本。一个 console script、一个进程内注册六个医学工具；initialize 和 tools/list 不读取模型参数或展开病例。第一次 `crc_lnm_predict_multimodal` 调用校验并加载唯一的 `seed_2024` NumPy runtime asset，后续预测复用同一实例。默认安装不依赖 PyTorch、NVIDIA 或 CUDA 包。
## ModelScope 正式配置
```json
{
"mcpServers": {
"crc-lnm-medical-agent": {
"command": "uvx",
"args": [
"crc-lnm-medical-agent@1.0.12"
]
}
}
}
```
ModelScope 中选择托管部署和 STDIO；command 填 `uvx`，args 只填上面一个带版本参数。不要增加 URL、host、port、transport 参数或环境变量。
## 本地验收
```powershell
powershell -ExecutionPolicy Bypass -File scripts/release_verify_full.ps1
```
六工具为模型信息、病例质控、CT 特征准备、病理特征准备、单模型预测和报告生成。内置病例仅为合成演示资源；本包只用于科研辅助，不构成诊断，所有输出均需专家复核。
TDQS
Scored across 2 tools
The two tools are clearly distinct: one reports health/status information, the other describes the deployment's behavior. There is no overlap in their purposes.
Naming styles are inconsistent: 'healthcheck' is a single word noun, while 'describe_deployment' follows the verb_noun snake_case pattern. This lack of a consistent convention makes the set feel uncoordinated.
With only two tools, the set is on the thin side, but the explicit mention of a 'deliberately limited behavior' suggests this small count is intentional for a staging environment. Still, it is borderline for a typical server.
Despite the server name indicating a medical agent, the provided tools only cover operational metadata (healthcheck and deployment description) with zero medical domain functionality. The surface is severely incomplete for any meaningful medical workflow.