ai-strategy
Pass
Audited by Gen Agent Trust Hub on Feb 20, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- Prompt Injection (LOW): The skill converts natural language descriptions into trading strategies, which is a significant surface for indirect prompt injection. 1. Ingestion points:
/strategyand/executecommand descriptions inSKILL.md. 2. Boundary markers: Absent; there is no clear separation between user instructions and data. 3. Capability inventory: Executing financial transactions on the Solana blockchain and monitoring price triggers. 4. Sanitization: No evidence of logic to validate trades or prevent instruction override in the provided interface. - No Code (SAFE): The skill consists only of a markdown definition file without accompanying source code. The security of how the
SOLANA_PRIVATE_KEYis handled or how transactions are signed and transmitted cannot be verified without the implementation logic.
Audit Metadata