email-imap-fetch
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external IMAP mailboxes which could contain malicious instructions designed to influence the behavior of the downstream agent or automation platform.
- Ingestion points: The
scripts/imap_idle_fetch.pyscript fetches email headers and content (subject and snippet) from configured IMAP accounts. - Boundary markers: The script wraps machine-readable metadata in delimiters (
<<<MAIL_REF_JSON>>>and<<<ATTACHMENT_MANIFEST_JSON>>>), but the emailsnippetandsubjectare passed as free-form text. - Capability inventory: The skill has network capabilities, specifically the ability to POST email data to external webhooks as defined in
scripts/imap_idle_fetch.py. - Sanitization: The script performs basic HTML tag removal and whitespace compaction in
extract_snippet, but does not implement specialized sanitization or "ignore embedded instructions" directives for the email content.
Audit Metadata