llm-trading-agent-security

Pass

Audited by Gen Agent Trust Hub on Aug 12, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill addresses the risk of untrusted data (token names, webhooks, social feeds) influencing agent behavior, which serves as an ingestion point for potentially malicious instructions.\n
  • Ingestion points: The sanitize_onchain_data function and general instructions highlight external data sources like webhooks or social feeds entering the LLM context in SKILL.md.\n
  • Boundary markers: The skill recommends architectural isolation, prompt hygiene, and independent spend policies as logical boundaries.\n
  • Capability inventory: The agent has the capability to sign and send on-chain transactions using eth-account and web3 providers.\n
  • Sanitization: Implements regex-based filtering (INJECTION_PATTERNS) to detect and block common injection strings such as instruction overrides or unauthorized transfer commands.\n- [EXTERNAL_DOWNLOADS]: The documentation references https://rpc.flashbots.net as a recommended RPC endpoint for MEV protection.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 12, 2026, 10:24 PM
Security Audit — agent-trust-hub — llm-trading-agent-security