Layer 2s

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

• Third-party content exposure detected (high risk: 0.95). The required runtime workflow uses curl -s https://ethskills.com/l2s/SKILL.md to fetch and ingest readable text from a public web page (outsider-authored content) into the agent’s LLM context.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The catalog explicitly instructs runtime fetching and use of external content—e.g., "curl -s https://ethskills.com/l2s/SKILL.md" and "claude plugin install https://github.com/austintgriffith/ethskills"—which will load remote content that directly controls agent instructions or installs/executes remote plugin code.