菜单
Expert guide for GCP deployment, security, and hospital-specific infrastructure. Covers Cloud Run, IAM, VPC, and secrets management for bioinformatics.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Audits all Markdown files in the precision-medicine-mcp repo against four canonical principles: (A) server Python code is the authoritative tool count, (B) server-registry.md is the authoritative server/tool summary, (C) all three patient outcomes (PAT001/PAT002/PAT003) must be represented consistently, (D) no content block may appear in more than one MD file (DRY). Use this skill whenever a new version of the platform ships, a new server is added, a new patient use case is validated, or before any major doc update. Trigger phrases: "audit the docs", "check docs are correct", "doc cleanup", "canonical check", "are the docs up to date", "DRY check on markdown".
Ensures all development meets clinical compliance and security standards. Focuses on HIPAA-friendly logging, audit trails, risk assessment, and scientific nomenclature.
Guide for managing patient data and bioinformatics datasets (h5ad, CSV). Standardizes data loading, preprocessing, and path management.
Expert guide for project-specific documentation and clinical reporting. Maintains consistency across persona-based guides and automated reports.
Expert guide for testing bioinformatics MCP servers. Covers pytest, DRY_RUN modes, and PatientOne simulation scenarios.
Expert guide for developing the Precision Medicine Streamlit dashboard and LLM providers. Covers provider abstraction, trace visualization, and UI components.
| name | precision-bio-infra |
| description | Expert guide for GCP deployment, security, and hospital-specific infrastructure. Covers Cloud Run, IAM, VPC, and secrets management for bioinformatics. |
This skill ensures that the platform is deployed securely and efficiently on Google Cloud or in clinical environments.
Standard deployment uses Cloud Run and Cloud Build:
roles/run.invoker and roles/storage.objectViewer.min-instances 0) to save costs, except for heavy-use multiomics servers.deploy.sh where provided (currently mcp-deepcell, mcp-cell-classify) and cloudbuild.yaml where provided (currently mcp-deepcell only).Follow the patterns in infrastructure/hospital-deployment/:
setup-vpc.sh to isolate genomic data traffic.deploy-oauth2-proxy.sh if deploying beyond IAM-restricted Cloud Run (e.g., GKE).setup-audit-logging.sh is run to track all tool invocations in Cloud Logging.shared/utils/bias_detection.py and the tools in infrastructure/audit/ when deploying new predictive models (e.g., GEARS).shared/utils/cost_tracking.py to ensure new infras include budget alerting.timeout appropriately (300s default, 600s for perturbation training).Use this skill when: