bingx-copytrade-spot
Pass
Audited by Gen Agent Trust Hub on Nov 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill consumes user-provided input that is subsequently used in authenticated API requests, which represents a potential surface for indirect prompt injection. \n- Ingestion points: The skill extracts parameters like orderId and timestamps from user queries within SKILL.md. \n- Boundary markers: The system requires the AI to wait for an explicit 'CONFIRM' command before executing trade operations on live accounts and mandates the use of specific signing logic. \n- Capability inventory: The skill enables sensitive actions including selling spot assets and querying private account details (profit, history) via the BingX API. \n- Sanitization: The implementation uses a validateParams function to reject URI-special characters and control characters, preventing parameter injection attacks. It also uses HMAC-SHA256 signatures to verify the integrity of the data sent to the API.
Audit Metadata