No red flags found in any of the 11 categories — no credential harvesting, no data exfiltration, no curl-pipe-shell installer. Scanned against the SlowMist agent-security taxonomy, refreshed every 8 hours. Full audit →
by charliehzm · MCP Server · ★ 95
Last updated: · Indexed by AgentSkillsHub · Auto-synced every 8h
🔒 Is medharness safe to install? View the security audit →
MedHarness · 医疗大模型流量安全与合规网关 给医院、医疗信息化厂商和医疗 AI 团队用的大模型统一出口： 每一次模型调用先过患者敏感信息识别、脱敏、分级路由、出站检查和审计留痕，再转发给允许的模型。 开源 · 支持私有化/专有云部署 · 面向中国医疗数据合规场景 （个人信息保护法 PIPL · 数据安全法 · 网络安全法/等保 · 健康医疗数据安全指南，兼顾 HIPAA） 说明：下文少量出现的 PHI 指患者健康医疗相关敏感信息，包括身份标识、就诊记录、检查检验、影像、诊断、用药等。 一句话：它是什么 HIS、EMR、CDR、互联网医院、随访系统、Dify/RAG/Agent，或者研发工具，只要把 指向 MedHarness， 每一次大模型调用都会先经过一条强制合规闸门： 识别患者敏感信息 → 脱敏 → 数据分级签名 → 模型白名单路由 → 提示注入扫描 → 出站安全检查 → 全量审计。 任一环节不确定或判定有风险，就直接拒绝，不连接上游模型。 它不是又一个通用大模型网关（LLM Gateway）。通用网关解决“接模型、限流、计费”，MedHarness 解决医疗场景里更关键的三件事： 患者数据能不能出、能出给哪个模型、出了事能不能查清楚。 HIS · EMR · CDR · 互联网医院 · Dify · RAG · Agent · 研发工具 │ OpenAI / Claude 兼容协议 —— 改 baseurl ▼ ┌──────────────────────────────────┐ │ MedHarness 合规网关 │ │ 识别→脱敏→分级→路由→注入→[转发]→出站→审计 │ └──────────────────────────────────┘ │ 脱敏后 · 白名单内 · 全链路留痕 ▼
| Stars | 95 |
| Forks | 9 |
| Language | Python |
| Category | MCP Server |
| License | Apache-2.0 |
| Quality Score | 60.1269319658675/100 |
| Open Issues | 11 |
| Last Updated | 2026-06-05 |
| Created | 2026-05-20 |
| Platforms | claude-code, mcp, python |
| Est. Tokens | ~1071k |
Looking for a medharness alternative? If you're comparing medharness with other mcp server tools, these 6 projects are the closest alternatives on Agent Skills Hub — ranked by topic overlap, star count, and community traction.
Code search MCP for Claude Code. Make entire codebase the context for any coding agent. Embeddings are created
OpenSpec Plus — Agentic skills that enhance OpenSpec's Spec-Driven Development through better discovery, requi
Open source AI coding platform with Web IDE, multi-agent system, 37+ tools, MCP protocol. MIT licensed.
Vendor-neutral MCP configuration manager — one CLI to add, toggle, and audit MCP servers and skills across eve
基于 Claude Code + OpenSpec + Superpowers 的全栈 AI 开发模板
OpenSkills: Run Claude Skills Locally using any LLM
Explore other popular mcp server tools:
medharness is 让医疗机构所有大模型流量：PHI 不外泄 · 模型走白名单 · 全量可审计 · 成本可控。. It is categorized as a MCP Server with 95 GitHub stars.
medharness is primarily written in Python. It covers topics such as ai-coding, claude-code, compliance.
You can find installation instructions and usage details in the medharness GitHub repository at github.com/charliehzm/medharness. The project has 95 stars and 9 forks, indicating an active community.
medharness is released under the Apache-2.0 license, making it free to use and modify according to the license terms.
The top alternatives to medharness on Agent Skills Hub include claude-context-local, openspec-plus, claude-code-open. Each offers a different approach to the same problem space — compare them side-by-side by stars, quality score, and community activity.
Grades come from a rule-based scan built on the SlowMist agent-security taxonomy, covering 11 red-flag categories including credential harvesting, data exfiltration, and curl | sh installers. It is a first-layer scan, not a manual audit — we say so rather than overstate it.
The scale of the problem is documented independently: Liu et al. (2026), in a study of 31,132 agent skills, report that 26.1% contain security vulnerabilities. Our own full-catalog census is published as a citable open dataset.
Sources & who's responsible: