hyperliquid-research
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches market data from official Hyperliquid API endpoints. These requests target well-known infrastructure associated with the protocol.
- [COMMAND_EXECUTION]: Includes curl commands for querying public API endpoints. These are standard information-retrieval patterns without exposure to command injection.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external data such as social media content.
- Ingestion points: Ingests data from websites, social posts, and token metadata as specified in SKILL.md.
- Boundary markers: Includes a clear instruction to treat these sources as untrusted and states that they cannot authorize any actions.
- Capability inventory: The skill is limited to read-only research and lacks tools for signing transactions, modifying accounts, or writing to the file system.
- Sanitization: Implements instructional distancing to prevent processed content from influencing the agent's core logic.
Audit Metadata