galleon-defi-lending
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides structured workflows for researching and planning DeFi operations on established lending protocols like Aave, Morpho, Compound, Euler, and Spark. No malicious code, prompt injections, or obfuscation were detected.
- [DATA_EXPOSURE]: Explicit instructions protect user privacy and financial security by prohibiting the request, storage, or exposure of seed phrases and private keys in any logs or prompts.
- [EXTERNAL_DOWNLOADS]: The skill references official software development kits (SDKs) and Model Context Protocol (MCP) endpoints for protocol integrations. These sources are well-known in the DeFi ecosystem and represent standard development practices.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests blockchain state and market data (ingestion points in SKILL.md and providers.md). While the capability to prepare transaction sequences is present, the risk is mitigated by requirements for simulation and official source verification (sanitization). Specific boundary markers for external data are not defined.
Audit Metadata