hyperliquid-monitor
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from the Hyperliquid API (e.g., order metadata, fill history, and account balances) to perform reconciliation.\n
- Ingestion points: External data sources include
frontendOpenOrders,userFillsByTime, andclearinghouseStateas defined inreferences/authoritative-state.md.\n - Boundary markers: The skill uses a structured markdown template defined in
references/output-contract.mdto constrain output format and mitigate injection risks.\n - Capability inventory: The skill is limited to read-only API queries and the execution of a bundled local mathematical script (
scripts/funding.mjs).\n - Sanitization: Instructions include explicit steps for deduplication, identifier validation, and multi-source state reconciliation to ensure data integrity.\n- [COMMAND_EXECUTION]: The skill executes a bundled Node.js script for localized mathematical calculations.\n
- Evidence: The script
scripts/funding.mjsis used to normalize funding intervals. It is a mathematical utility that operates entirely on provided arguments without external network or filesystem dependencies.
Audit Metadata