auditing-wallet-security

Pass

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill is subject to an indirect prompt injection surface through the processing of untrusted blockchain data.\n
  • Ingestion points: The approval_scanner.py script fetches token names and symbols via RPC calls, and tx_analyzer.py retrieves contract information from block explorer APIs.\n
  • Boundary markers: There are no explicit delimiters or specific instructions for the agent to distinguish between legitimate data and potential instructions embedded within blockchain metadata.\n
  • Capability inventory: The skill has the ability to execute its internal Python scripts via Bash(crypto:wallet-*) and manage files through Read, Write, and Edit tools.\n
  • Sanitization: While wallet_auditor.py validates the format of the input wallet address, the skill does not implement sanitization for the arbitrary text strings (token names, symbols) retrieved from external sources.\n- [EXTERNAL_DOWNLOADS]: The skill references and fetches data from well-known technology services and blockchain infrastructure providers.\n
  • It makes HTTP requests to established block explorer APIs including Etherscan, BSCScan, PolygonScan, Arbiscan, and BaseScan to query transaction logs and contract source code status.\n
  • It utilizes public RPC endpoints for multiple blockchains, including providers like LlamaNodes and Binance's dataseed, to retrieve on-chain events and perform contract calls.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 12, 2026, 01:12 AM