bingx-coinm-ws-account
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes real-time account and trade data from a WebSocket stream, creating a surface for potential indirect prompt injection attacks. 1. Ingestion points: The
ws.onmessagehandler inSKILL.mdreceives and processes external data from the BingX exchange. 2. Boundary markers: The skill includes 'Agent Interaction Rules' that explicitly prohibit the agent from returning code blocks to the user and restrict behavior to GET-only execution. 3. Capability inventory: The skill usesfetchfor authentication and maintains a persistent WebSocket connection. 4. Sanitization: Input data is decompressed via GZIP and parsed as JSON, and the agent is instructed to validate all parameters. - [EXTERNAL_DOWNLOADS]: The skill makes network requests to official BingX API and WebSocket endpoints. These are identified as legitimate vendor resources for the stated purpose of account monitoring.
Audit Metadata