indexing

Warn

Audited by Snyk on Feb 19, 2026

Risk Level: MEDIUM
Full Analysis

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

  • Third-party content exposure detected (high risk: 0.90). Yes — SKILL.md explicitly instructs querying public third-party sources (The Graph subgraphs via studio.thegraph.com, Alchemy/QuickNode/Etherscan APIs, Dune, and RPC/WebSocket event logs) and expects the agent to read and act on those on-chain/subgraph/API results, which are public/untrusted content that can materially influence actions.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 19, 2026, 07:10 PM