linkfox-shopee-store-orders
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The helper script
scripts/_shopee_orders_common.pyusessubprocess.run()to execute a local dependency checker script (check_auth_dependency.py). This is used for internal validation of required skills and does not involve untrusted input. - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from the Shopee Open API and stores it in local JSON files for the agent to process. This creates an attack surface where malicious content embedded in Shopee order data (e.g., in order notes or item names) could potentially influence the agent's behavior.
- Ingestion points: Shopee API responses are fetched via the
developerProxyand processed inscripts/_shopee_orders_common.py, then saved to files via theemit_result()function. - Boundary markers: Not present. Data is saved as raw JSON strings for agent consumption.
- Capability inventory: Includes file system writing (
emit_result), network requests to vendor-controlled domains (call_api), and local script execution (subprocess.run). - Sanitization: Standard JSON deserialization is performed, but no specific sanitization or filtering is applied to the content of the API response before it is presented to the agent context.
Audit Metadata