lavarage
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill communicates with the official Lavarage API (https://api.lavarage.xyz) to fetch market offers, token metadata, and build Solana transactions. This is essential for the skill's primary function of facilitating leveraged trading.
- [COMMAND_EXECUTION]: The example scripts demonstrate the deserialization, signing, and submission of Solana transactions using @solana/web3.js. These are intended capabilities for executing on-chain trades and do not include arbitrary shell command execution.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data such as token names, symbols, and descriptions fetched from the Lavarage API. This presents an attack surface where a malicious actor could theoretically inject instructions via token metadata, though the risk is localized to the trading context.
- Ingestion points: Metadata is ingested via getTokens and getOffers in templates/client.ts and examples/discover-markets.ts.
- Boundary markers: None identified.
- Capability inventory: The skill has the capability to sign and submit financial transactions to the Solana blockchain.
- Sanitization: The client library does not perform explicit sanitization of token metadata strings before they are processed by the agent.
Audit Metadata