ae-handoff-brief-skill
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from multiple external platforms, creating an attack surface for indirect prompt injection.
- Ingestion points: The skill ingests data from HubSpot contact records, Clari call transcripts via
clari_get_call_summary, and Apollo lead enrichment viaenrich_contactas defined in the workflow stages 2a, 2b, and 2c. - Boundary markers: The skill utilizes a predefined markdown template in
reference/ae-handoff-template.mdto structure the output, but does not implement robust delimiters to isolate potentially malicious text within transcripts or contact data. - Capability inventory: The agent has the capability to write HubSpot notes using
hubspot_create_contact_noteand create Gmail drafts usingGmail.create_draft, providing a path for injected instructions to influence downstream communication or CRM records. - Sanitization: While the skill includes a brand consistency check via
check_my_copy(Gate B), it lacks explicit security filtering or sanitization for external content before interpolation into the prompt for the final brief.
Audit Metadata