podium-multi-location-router
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements strict per-location credential isolation, ensuring that OAuth tokens and session state for one physical location are never reused or shared with another, even within the same organization.
- [DATA_EXFILTRATION]: Network communication is exclusively directed to official Podium API and authentication domains. The structured audit logging system is purpose-built to record routing metadata while strictly excluding sensitive information such as API tokens, request bodies, or PII.
- [SAFE]: Secure file system operations are used throughout, including atomic writes (via temporary files and renaming) and the enforcement of restrictive file permissions (mode 0600) for sensitive credential maps.
- [INDIRECT_PROMPT_INJECTION]: The skill incorporates robust boundary verification by performing pre-flight checks against the Podium API (
/v4/locations) before executing any write operations. This prevents the agent from being induced to perform unauthorized actions on incorrect or malicious location IDs provided in external data. - [COMMAND_EXECUTION]: External tool usage (python3, sqlite3, jq, curl) is limited to standard administrative and data-querying tasks as defined in the technical architecture documentation.
Audit Metadata