claude-agent-sdk

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly exposes WebSearch/WebFetch in SKILL.md ("Built-in Tools" and the "Web-augmented" recommendation) and includes examples in references/custom-tools.md that fetch and parse external URLs (e.g., fetch to https://api.example.com), so the agent is expected to ingest and act on untrusted public web content which could influence tool use and decisions.