synthesis-inbox-cleanup

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

• Third-party content exposure detected (medium risk: 0.30). At runtime, iCloud/IMAP paths such as scripts/icloud_inspect_senders.py fetch attacker-controlled email bodies from the user’s INBOX and then route the sanitized <UNTRUSTED_EMAIL nonce="...">...</UNTRUSTED_EMAIL> text into an LLM-assisted triage boundary, creating an indirect prompt-injection exposure surface.