Weather Automation

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [SAFE]: No malicious patterns or security vulnerabilities were identified in the skill configuration.
  • [NO_CODE]: The skill consists entirely of configuration and documentation (YAML and Markdown). No executable scripts or binary files are provided, significantly reducing the attack surface.
  • [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection as it processes external data from weather services and calendar events. This is a common design pattern for automation tools.
  • Ingestion points: Weather data from weather-mcp tools and calendar event metadata in SKILL.md.
  • Boundary markers: None identified in the workflow templates.
  • Capability inventory: Actions include fetching weather data and sending notifications via Slack, SMS, or Home Assistant.
  • Sanitization: The templates do not show explicit input sanitization or validation of the external data before use.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 05:18 AM