copy-trading
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from multiple external sources (SolanaTracker, Helius, DexScreener) to drive its evaluation and monitoring logic. This creates an indirect prompt injection surface where the agent could be influenced by malicious data provided by these third-party services.
- Ingestion points: Data is fetched via 'fetch_wallet_pnl' in 'scripts/evaluate_wallet.py', and 'fetch_recent_signatures'/'fetch_token_info' in 'scripts/monitor_wallet.py'.
- Boundary markers: The scripts parse structured JSON but do not implement explicit boundary markers or instructions to ignore potential natural language commands that could be embedded in token metadata or transaction logs.
- Capability inventory: The skill description includes instructions for automated trade execution via the Jupiter aggregator based on the parsed data.
- Sanitization: No specific sanitization or validation of the fetched data for natural language injection patterns is implemented.
Audit Metadata