uptimerobot-automation

Pass

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references the Rube MCP server endpoint at https://rube.app/mcp and documentation at composio.dev. These are established resources for the Composio automation platform.
  • [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection because it processes tool schemas and execution plans dynamically retrieved from a remote source.
  • Ingestion points: Data entering the agent via RUBE_SEARCH_TOOLS responses which contain dynamic tool slugs and argument schemas.
  • Boundary markers: None identified; instructions do not explicitly mandate delimiters for external tool data.
  • Capability inventory: The skill uses RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH to perform operations based on external inputs.
  • Sanitization: None identified; the skill relies on the agent's interpretation of the discovered schemas.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 29, 2026, 07:07 AM