quicknode-cost-tuning
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns detected. The skill provides legitimate implementation patterns for QuickNode blockchain integration.
- [INDIRECT_PROMPT_INJECTION]: The skill interacts with external blockchain RPC endpoints. While the provided snippet only logs a block number, reading data from external APIs/RPCs introduces a theoretical surface for indirect prompt injection if the agent uses the data to drive autonomous actions.
- Ingestion points: Data retrieved via
ethers.JsonRpcProvider(e.g.,getBlockNumber). - Boundary markers: None provided in the code snippet.
- Capability inventory: The skill allows
Write,Edit, andBashtools. - Sanitization: No explicit sanitization or validation of RPC responses is shown in the boilerplate snippet.
Audit Metadata