gateway-diagnose

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

• Third-party content exposure detected (high risk: 0.70). The skill explicitly ingests and inspects user-generated Telegram messages and session transcripts (see channels/telegram.ts and the ~/.joelclaw/sessions/gateway/*.jsonl and gateway logs) as part of its required diagnostic workflow, which are untrusted third‑party inputs that can influence decisions and next actions.