ecommerce-geekbi-temu-shop
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill makes network requests to
api.nexscope.aito retrieve ecommerce data. These requests are directed to the vendor's official API gateway and are necessary for the skill's primary research functions. - [COMMAND_EXECUTION]: The skill executes Python scripts (
geekbi_temu_category_list.py,geekbi_temu_mall_search.py,geekbi_temu_site_list.py) to wrap API calls. These scripts use standard libraries (e.g.,urllib,json,os) and do not perform unauthorized shell commands or system modifications. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the Temu marketplace (e.g., shop names, product descriptions).
- Ingestion points: Data enters the agent context through the business response object in all three utility scripts.
- Boundary markers: The scripts do not implement explicit delimiters for the summarized output of retrieved data.
- Capability inventory: The skill has network access to the vendor API and local file-writing capabilities for logging.
- Sanitization: The skill parses JSON data but does not perform specific sanitization for prompt injection characters before summarization, though the risk is considered low given the data-retrieval context.
- [SAFE]: The skill follows security best practices by reading credentials from environment variables, explicitly warning against the exposure of raw secrets, and limiting operations to public-market research.
Audit Metadata