defi-attack-patterns

Pass

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill defines a decision tree for analyzing external DeFi protocols, which involves ingesting untrusted data such as smart contract code and protocol parameters. This creates a potential surface for indirect prompt injection if the analyzed content contains malicious instructions.
  • Ingestion points: Section 9 ('Decision Tree') directs the agent to evaluate external protocol components during analysis.
  • Boundary markers: Absent. There are no instructions to wrap or delimit analyzed protocol data to prevent the agent from executing instructions embedded within it.
  • Capability inventory: The skill focus is instructional analysis; it does not explicitly invoke shell commands, file writes, or network operations within the SKILL.md content.
  • Sanitization: Absent. The instructions do not include steps to sanitize or validate the external protocol data before it is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 21, 2026, 01:24 PM
Security Audit — agent-trust-hub — defi-attack-patterns