The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted product data such as names, descriptions, and attributes from Temu APIs. This data enters the agent context via scripts like global_manage_detail_query.py without explicit sanitization, validation, or boundary markers. While no specific malicious payload was found, the capability to read and process external data creates an attack surface for indirect prompt injection.
[DATA_EXFILTRATION]: Network operations are directed towards tool-gateway.linkfox.com, api.linkfox.com, and agent-api.linkfox.com. These domains are owned by the vendor, linkfox-ai, and are used for API proxying, user authentication, and billing management. No unauthorized data exfiltration to third-party domains was detected.
[COMMAND_EXECUTION]: The skill operates by executing internal Python scripts to perform business logic and API requests. This is the intended behavior for the tool's functionality.
[EXTERNAL_DOWNLOADS]: The onboarding.py utility includes instructions for users to manually install standard Python libraries like requests, qrcode, and pillow if they are missing.
[DATA_EXPOSURE]: The skill facilitates the local storage of Temu access tokens in the user's home directory at ~/.linkfox/temu-access-tokens.json to enable reuse across sessions. This is a common practice for CLI credentials management but represents a local exposure of those specific secrets.

linkfox-temu-manage-product-global