hyperliquid-websocket
Pass
Audited by Gen Agent Trust Hub on Aug 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a process for ingesting external data that could potentially contain malicious instructions.\n
- Ingestion points: Data is received via WebSocket streams from api.hyperliquid.xyz/ws (e.g., trades, market mids, and user events).\n
- Boundary markers: The skill does not provide markers or instructions to isolate the ingested data from the agent's command context.\n
- Capability inventory: The skill includes Python code that writes incoming data to a local log file (/workspace/trading-desk/watch/ws.log) and outputs it to the console.\n
- Sanitization: There is no evidence of filtering or sanitizing the streamed JSON data before it is logged or displayed.
Audit Metadata