check-vault-onchain

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMCOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION] (MEDIUM): The skill instructions explicitly direct the agent to run a Python script using the python console command. This involves executing code on the underlying system.
  • [INDIRECT_PROMPT_INJECTION] (LOW): The skill processes user-provided inputs such as blockchain explorer links and smart contract addresses.
  • Ingestion points: User-provided URLs or blockchain addresses.
  • Boundary markers: None specified in the instructions.
  • Capability inventory: Execution of a Python script (scripts/erc-4626/check-vault-onchain.py) and potential web browsing to fetch contract data.
  • Sanitization: No sanitization or validation of the input strings/links is mentioned before they are incorporated into the script or process.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 06:15 PM