qinghu-amazon-keyword-picker
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill initiates network requests to an external API endpoint at
https://www.iqinghu.com/api/desktop/qh/mcp. It instructs the agent to read sensitive authentication tokens from environment variables (QINGHU_TOKEN,QHKIT_TOKEN) or prompt the user for them to authorize these requests. These operations are core to the skill's primary function of retrieving marketplace data from the vendor's platform. - [INDIRECT_PROMPT_INJECTION]: The skill acts as a consumer of data from external search and research tools provided by the vendor API, creating a vulnerability surface where malicious instructions could be embedded in the retrieved data.
- Ingestion points: Data is ingested through tool results (e.g.,
keyword_miner,product_research) parsed from thecontent[0].textfield of the API response. - Boundary markers: There are no instructions to use specific delimiters or to treat the external tool output as untrusted content.
- Capability inventory: The skill has the capability to perform network operations and can trigger external file export skills like
qinghu-excel-export. - Sanitization: The skill instructions do not specify any validation, filtering, or escaping mechanisms for the external data before it is processed or presented to the user.
Audit Metadata