bio-ai-product-manager

Pass

Audited by Gen Agent Trust Hub on Apr 22, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is composed entirely of informational Markdown documents and templates for drafting product requirements. There is no executable code or script logic included in the skill folder.\n- [COMMAND_EXECUTION]: There are no shell commands, system operations, or instructions for the agent to execute code on the host system.\n- [DATA_EXFILTRATION]: No network activity or commands to access sensitive directories (e.g., .ssh, .aws) were found. The documentation explicitly advises against transmitting Protected Health Information (PHI) to external models.\n- [PROMPT_INJECTION]: The instructions do not utilize techniques to override agent constraints or bypass safety protocols. Input handling is focused on domain-specific requirements analysis.\n- [CREDENTIALS_UNSAFE]: No hardcoded secrets, API keys, or tokens are present. References to API endpoints (e.g., CCDI Federation API) are used in a purely descriptive and illustrative context.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 22, 2026, 06:24 PM