linkfox-shopee-store-public
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill handles data from an external platform which introduces an indirect prompt injection surface.
- Ingestion points: Shopee Open API responses are fetched via the developer proxy in
scripts/_public_api_runner.pyand written directly to stdout or session JSON files in the workspace, which are subsequently parsed by the agent. - Boundary markers: The output does not use specific delimiters or instructions to ignore potential natural language commands embedded in the Shopee response data.
- Capability inventory: The skill is capable of performing HTTPS requests via
urllib.request.urlopento the proxy endpoints and triggering controlled local subprocess calls. - Sanitization: The Shopee response objects are handled as structured data or raw text without specialized sanitization or escaping filters before context ingestion.
- [COMMAND_EXECUTION]: The skill invokes an external process to verify dependency availability.
- Evidence: In
scripts/_shopee_public_common.py,subprocess.runis used to executecheck_auth_dependency.pysafely usingsys.executablewith arguments passed as a list. It does not spawn a shell, avoiding shell injection risks.
Audit Metadata