AgentSkillsCN

security-baseline

适用于整个系统的安全需求、潜在威胁及控制措施。

SKILL.md
--- frontmatter
name: security-baseline
description: Security requirements, threats, and controls that apply across this system.

Security Baseline

Threat Model (High Level)

  • •Primary users: internal analytics users
  • •Assets to protect:
    • •Customer PII (anonymized in demo)
    • •Event data integrity
    • •Connection strings and secrets

Required Controls

Authentication

  • •Use Azure AD / Entra ID for Fabric workspace access
  • •Service principals for automated pipelines

Authorization

  • •Enforce least privilege on Fabric items
  • •Separate dev/prod workspaces

Input Validation & Output Encoding

  • •Validate all untrusted input at boundaries
  • •Sanitize or encode output where appropriate

Secrets Management

  • •Store secrets in Azure Key Vault, never in code or config files
  • •Use Fabric-managed connections where possible

Logging & Auditing

  • •Log security-relevant events
  • •Avoid logging sensitive data

Common Vulnerabilities

  • •Never embed connection strings in notebooks
  • •Validate JSON payloads in streaming pipelines