linkfox-temu-returns-refunds-global
Pass
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: SAFEDATA_EXFILTRATIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it processes data from external API responses.
- Ingestion points: API response bodies ingested from
tool-gateway.linkfox.comvia scripts liketemu_proxy.pyandtemu_global_proxy.py. - Boundary markers: Absent; the skill does not wrap external API content in delimiters or provide explicit warnings to the agent to ignore instructions within the data.
- Capability inventory: Network access via
urllibandrequests, file system writes to the workspace and home directory, and execution of multiple Python scripts. - Sanitization: No sanitization or filtering is performed on the content of the API responses beyond JSON parsing.
- [DATA_EXFILTRATION]: The skill manages sensitive credentials by reading from and writing to
~/.linkfox/temu-access-tokens.jsonto persist Temu access tokens. Additionally, it logs full API request/response data to the project workspace under thelinkfox/directory. - [COMMAND_EXECUTION]: The skill includes a variety of Python scripts meant to be executed by the agent to proxy requests to the Temu platform. It also includes an onboarding script that facilitates account management and login operations.
Audit Metadata