ecommerce-tiktok-product-search
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill retrieves product metadata (names, titles, and categories) from TikTok Shop via the EchoTik API. This data is ingested into the agent's context without sanitization or explicit boundary markers, creating a surface for potential indirect prompt injection if an attacker crafts malicious product listings.
- Ingestion points: Data is fetched via
scripts/echotik_list_product.pyand presented to the agent. - Boundary markers: No delimiters or 'ignore embedded instructions' warnings are present in the script output or SKILL.md instructions.
- Capability inventory: The skill possesses file-write capabilities (saving responses to the
nexscope/directory) and network access (urllib.request.urlopen). - Sanitization: The script does not sanitize or escape the retrieved product data before it is printed to stdout or saved to disk.
- [COMMAND_EXECUTION]: The skill includes a standalone Python script (
scripts/echotik_list_product.py) that performs network requests and writes JSON data to the local filesystem within the current working directory. While these operations are functional requirements for the tool, they represent broad file system and network capabilities assigned to the agent.
Audit Metadata