← Back to Skills Marketplace
gptplusplus

Medical Science

by 船长 · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ pending
30
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install medical-science
Description
Provides integrated tools for clinical trial monitoring, medical writing, competitor analysis, and knowledge management with full regulatory compliance.
README (SKILL.md)

Medical Science Skills — 医药研发技能集

Production-ready skills for clinical trial monitoring, medical writing, competitive intelligence, and knowledge management.

Metadata

Name: Medical Science Skills (医药研发技能集) Version: 1.0.0 Category: Medical / Healthcare Author: Medical Science Team License: MIT

Description

面向医药研发全流程的 OpenClaw 技能包。覆盖临床试验监测(安全性/疗效/受试者画像/方案偏离/入排核查/综合仪表盘)、医学撰写(Protocol/CSR/eCTD/SAP)、竞品调研(月度报告/靶点分析)、知识管理(文献阅读/Wiki 加工/项目交接)四大领域,共 15 个技能。所有技能共享统一 MySQL 数据库,遵循 GCP (ICH E6 R3)、ICH E3/E9/M11、CTCAE v5.0、MedDRA、21 CFR Part 11、ALCOA+ 等合规标准。

Skills Included (15)

医学监察 (Monitoring) — 6 skills

  1. monitoring-safety — 安全性监测:AE/SAE/CTCAE 迭代分析,MedDRA SOC/PT 信号识别,Hy's Law 检测,DSMB 报告
  2. monitoring-efficacy — 疗效监测:主要/次要终点追踪,访视轴折线图,RECIST 响应分布,ORR/PFS 统计
  3. monitoring-patient-profile — 受试者画像:全维度时间轴(药物暴露/访视/AE/实验室/病史/合并用药)
  4. monitoring-pd-detector — 方案偏离检测:访视超窗/ICF 版本错误/关键检查缺失自动识别,CAPA 关联
  5. monitoring-enrollment-check — 入排核查:方案入排 vs 受试者资料比对,合规性判断与原始资料参考
  6. monitoring-dashboard — 综合仪表盘:疗效/安全性/实验室/受试者状态四维预警视图

医学撰写 (Writing) — 4 skills

  1. writing-protocol — 临床方案撰写:ICH E6(R3)/E8/SPIRIT 2013 标准,18 章节结构
  2. writing-csr — CSR 辅助撰写:ICH E3 标准,TFL → 自然语言,数据一致性校验
  3. writing-ectd — eCTD 文档导出:M1-M5 完整性核查,XML Schema/序列号/MD5 校验
  4. writing-statistics — 统计分析计划 (SAP):ICH E9 标准,样本量计算,分析人群定义

竞品调研 (Competitor) — 2 skills

  1. competitor-report — 月度竞品报告:ClinicalTrials.gov 数据整合,里程碑追踪,结构化月报
  2. competitor-target — 靶点竞争格局:竞争强度评估,进度差距分析,空白适应症识别

知识管理 (Knowledge) — 3 skills

  1. knowledge-literature — 文献智能阅读:PICO 提取,JADAD/ROB/GRADE 质量评估,结构化笔记
  2. knowledge-wiki — Wiki 知识加工:结构化页面,知识图谱节点提取,冲突检测
  3. knowledge-handover — 知识交接:Baton 接力棒系统,跨会话/跨成员项目知识传递

Key Benefits

  • 全流程覆盖 — 从临床前竞品调研到 NDA 提交,15 个技能覆盖医药研发核心环节
  • 合规内建 — GCP/ICH/CTCAE/MedDRA/21 CFR Part 11/ALCOA+ 合规标准嵌入每个技能
  • 统一数据层 — 45+ 张表的 MySQL 数据库,所有技能共享,避免数据孤岛
  • 触发词驱动 — 自然语言触发,无需记忆命令(如输入"AE/SAE"即触发安全性监测)
  • 结构化输出 — 每个技能输出标准化报告(Markdown/CSV/XLSX/PDF),可直接用于监管提交

Quick Start

# 安装技能包
openclaw skills install medical-science-skills

# 配置数据库连接(编辑技能目录下的 .env 文件)
cd skills/medical-science
cp .env.example .env   # 如有模板则复制,否则新建 .env,内容见下方
# 编辑 .env 填入 DB_HOST / DB_PORT / DB_NAME / DB_USER / DB_PASSWORD

# 初始化数据库(db.js 会自动读取 .env 中的配置)
node scripts/db.js --file database/schema.sql

配置方式: 技能根目录的 .env 文件(已自动被 db.js 加载),优先级:显式 export 环境变量 > .env 文件 > 硬编码默认值。 安全提示: .env 已在 .gitignore 中,不会被提交到版本控制系统。

Scripts (Node.js 辅助工具)

技能包内置 Node.js 辅助脚本,用于复杂计算和外部 API 集成:

脚本 功能 使用技能
scripts/db.js 数据库连接池(统一入口) 全部
scripts/signal-detection.js PRR/ROR/BCPNN 安全信号统计算法 monitoring-safety
scripts/meddra-lookup.js MedDRA 层级遍历(SOC→HLGT→HLT→PT) monitoring-safety
scripts/recist-engine.js RECIST 1.1 规则引擎(CR/PR/SD/PD 判定) monitoring-efficacy
scripts/km-curve.js Kaplan-Meier 生存分析 + Log-rank 检验 monitoring-efficacy
scripts/dose-calculator.js 剂量强度/累积剂量/调整率计算 monitoring-patient-profile
scripts/kri-compute.js RBQM 关键风险指标 + 中心风险评分 monitoring-dashboard
scripts/sample-size.js 样本量计算(t检验/比例/优效/非劣效) writing-statistics
scripts/ectd-validator.js eCTD XML 校验 + MD5/SHA256 计算 writing-ectd
scripts/ctgov-api.js ClinicalTrials.gov API v2 客户端 competitor-report, competitor-target
scripts/pubmed-api.js PubMed E-utilities 客户端 knowledge-literature
scripts/report-generator.js PDF/XLSX 报告生成 全部输出报告的技能

脚本统一从环境变量读取数据库配置(.envprocess.env → 硬编码默认值),使用方式:node scripts/xxx.js --help

Requirements

  • Node.js 18+
  • MySQL 8.0+
  • 数据库环境变量(DB_HOST / DB_PORT / DB_NAME / DB_USER / DB_PASSWORD)
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install medical-science
  3. After installation, invoke the skill by name or use /medical-science
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Medical Science Skills v1.0.0 — initial production-ready release - Introduces a comprehensive skill set (15 total) for clinical trial monitoring, medical writing, competitor analysis, and knowledge management. - Ensures compliance with major medical research standards (GCP/ICH/CTCAE/MedDRA/21 CFR Part 11/ALCOA+). - Unified MySQL database structure supports all skills, reducing data silos. - Includes Node.js helper scripts for statistical computation, data validation, and API integration. - Designed for structured outputs suitable for regulatory use (Markdown/CSV/XLSX/PDF). - Natural language triggers streamline user experience—no command memorization needed.
Metadata
Slug medical-science
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is Medical Science?

Provides integrated tools for clinical trial monitoring, medical writing, competitor analysis, and knowledge management with full regulatory compliance. It is an AI Agent Skill for Claude Code / OpenClaw, with 30 downloads so far.

How do I install Medical Science?

Run "/install medical-science" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Medical Science free?

Yes, Medical Science is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Medical Science support?

Medical Science is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Medical Science?

It is built and maintained by 船长 (@gptplusplus); the current version is v1.0.0.

💬 Comments