linkfox-temu-order-global
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONCREDENTIALS_UNSAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill executes various Python scripts to interact with Temu APIs and manage store-specific configurations.
- [DATA_EXFILTRATION]: API keys and access tokens are transmitted to vendor-controlled endpoints (tool-gateway.linkfox.com, api.linkfox.com, and agent-api.linkfox.com) to facilitate authenticated business operations. This activity is restricted to the infrastructure of the skill author, linkfox-ai.
- [CREDENTIALS_UNSAFE]: The skill maintains Temu access tokens in a local JSON file located at ~/.linkfox/temu-access-tokens.json for persistent store management.
- [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection via external API data ingestion.
- Ingestion points: The skill retrieves order lists, details, customization text, and shipping information from Temu APIs in scripts like global_order_list_v2_get.py and global_order_detail_v2_get.py.
- Boundary markers: No delimiters or safety instructions are used to separate retrieved data from the agent's internal instruction context.
- Capability inventory: The skill scripts have network access and the ability to read/write local session data and configuration files.
- Sanitization: External data fields, which may include customer-supplied text, are parsed and presented to the agent without sanitization or filtering.
Audit Metadata