sn-ppt-workbench
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONCREDENTIALS_UNSAFEDATA_EXFILTRATIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes 'node' via 'subprocess.run' in Python and 'child_process.spawn' in Node.js to launch the workbench launcher and server components. These commands use argument lists to mitigate shell injection risks.
- [CREDENTIALS_UNSAFE]: The skill manages authentication tokens, such as 'API_SERVER_KEY' and 'WORKBENCH_GATEWAY_API_KEY', by reading from vendor-specific configuration files (e.g., '~/.hermes/.env'). This follows standard environment-based secret management practices.
- [DATA_EXFILTRATION]: The Python helper initiates a network request to a 'Gateway' URL (defaulting to whitelisted localhost) to verify session connectivity. This request includes a Bearer token used for bridging agent session data to the workbench.
- [DYNAMIC_EXECUTION]: The Node.js launcher script spawns the workbench server as a background process to provide the interactive WebUI for slide editing.
Audit Metadata