breezy-hr-automation

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [NO_CODE] (SAFE): The skill consists entirely of markdown documentation. No scripts (Python, JavaScript, Shell) or binaries are included in the package.
  • [EXTERNAL_DOWNLOADS] (SAFE): The skill references the Rube MCP endpoint (https://rube.app/mcp) and Composio documentation. These are standard configuration references for the intended functionality and do not involve automated or hidden downloads of untrusted code.
  • [DATA_EXFILTRATION] (SAFE): No patterns of credential harvesting or unauthorized data transmission were detected. Authentication is handled via the documented RUBE_MANAGE_CONNECTIONS tool pattern which redirects users to an official OAuth/auth flow.
  • [INDIRECT_PROMPT_INJECTION] (LOW): The skill is designed to process data from an external HR system (Breezy HR).
  • Ingestion points: Breezy HR candidate data and job descriptions retrieved via tool outputs.
  • Boundary markers: Absent; the instructions do not specify the use of delimiters for external data.
  • Capability inventory: Uses RUBE_MULTI_EXECUTE_TOOL which can perform write actions in the HR system.
  • Sanitization: Not explicitly documented.
  • Assessment: While an attack surface exists if an attacker places malicious instructions in a job application, the risk is inherent to the HR automation use case and handled by the underlying LLM's safety filters.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:34 PM