euler-irm-oracles

Warn

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [External Downloads] (MEDIUM): The skill encourages the use of external scripts from the euler-xyz GitHub organization (e.g., calculate-irm-linear-kink.js and calculate-irm-adaptive-curve.js). Although these are official protocol tools, the organization is not within the predefined trusted scope (e.g., vercel-labs, anthropics), rendering the dependencies unverifiable according to the strict security policy.
  • [Indirect Prompt Injection] (MEDIUM): The skill defines a workflow for ingesting external oracle data to influence high-impact protocol configurations.
  • Ingestion points: Price feed data from Chainlink, Pyth (Hermes API), and Chronicle.
  • Boundary markers: None (data is processed as raw numeric or bytes values).
  • Capability inventory: Configuring the EulerRouter via govSetConfig and govSetResolvedVault, and deploying IRM contracts via factories.
  • Sanitization: Includes best-practice staleness checks (maxStaleness) and confidence interval validation for Pyth, but remains vulnerable to manipulated feed data that could influence the agent's financial logic.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 16, 2026, 02:25 AM