desk-execution-protocol
Pass
Audited by Gen Agent Trust Hub on Aug 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data which could potentially contain adversarial instructions if the source is compromised.
- Ingestion points: Reads trade parameters from
/workspace/trading-desk/proposals/<id>.mdand consumes live market data/order statuses from Hyperliquid API endpoints (allMids,meta,orderStatus,clearinghouseState). - Boundary markers: The protocol requires a 'Risk Manager PASS' metadata tag within the proposal file and a corresponding 'user's approval line' in the chat history to validate the intent.
- Capability inventory: The skill has the capability to perform financial transactions (orders, cancels, leverage changes) via the Hyperliquid exchange API.
- Sanitization: The skill implements robust data validation, including rounding prices to specific significant figures, rounding sizes down to decimal limits, slippage tolerance checks, and trigger price side validation before any action is taken.
Audit Metadata