linkfox-temu-promotion-global
Pass
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: SAFECREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill has an indirect prompt injection surface due to the ingestion of data from external API responses. \n
- Ingestion points: Data enters the context through responses from the Temu Global Promotion APIs via the LinkFox gateway (e.g.,
scripts/temu_global_proxy.py). \n - Boundary markers: The skill does not implement specific delimiters or instructions to prevent the agent from interpreting instructions potentially embedded in API data. \n
- Capability inventory: The skill can perform network operations to LinkFox domains and write data to the local file system (e.g.,
linkfox/result directory). \n - Sanitization: API responses are handled as JSON objects but lack filtering to sanitize natural language content. \n- [EXTERNAL_DOWNLOADS]: The skill makes network requests to external domains associated with the vendor's infrastructure. \n
- Communicates with
tool-gateway.linkfox.comfor API proxy operations. \n - Communicates with
api.linkfox.comandagent-api.linkfox.comfor authentication, user management, and billing services. \n- [COMMAND_EXECUTION]: The skill provides various Python scripts to be executed via the CLI for API interaction and token management. \n- [CREDENTIALS_UNSAFE]: The skill includes mechanisms for managing authentication tokens. \n - Stores Temu access tokens locally in
~/.linkfox/temu-access-tokens.jsonto provide session persistence. \n - The
onboarding.pyscript provides instructions for the user to set theirLINKFOXAGENT_API_KEYas a system environment variable for gateway authentication.
Audit Metadata