qinghu-tiktok-bluesea-collector
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFE
Full Analysis
- [DATA_EXPOSURE]: The skill manages API authentication by instructing the agent to look for 'QINGHU_TOKEN' or 'QHKIT_TOKEN' environment variables or to request a key from the user. This follows standard practices for secret management in agent skills.
- [EXTERNAL_DOWNLOADS]: The skill communicates with 'www.iqinghu.com' to retrieve TikTok product rankings, search results, and 1688 supply chain data. These network operations are intrinsic to the skill's stated purpose of e-commerce data collection.
- [COMMAND_EXECUTION]: The instructions suggest using available system tools like 'curl' or programming libraries like 'requests' or 'fetch' to interact with the Qinghu JSON-RPC API endpoints if native agent tools are not pre-configured.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external content including TikTok product titles, video metadata, and 1688 supplier descriptions. This creates a surface for indirect prompt injection where untrusted external data enters the agent context.
- Ingestion points: API responses from TikTok and 1688 data tools in 'SKILL.md'.
- Boundary markers: The skill relies on JSON structure to separate data from instructions.
- Capability inventory: The agent can perform network requests and parse nested JSON data structures.
- Sanitization: The instructions focus on JSON parsing and error handling but do not specify content sanitization for the retrieved text strings.
Audit Metadata