account-handover

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes untrusted data from Vitally notes and conversations to synthesize handover documents.
  • Ingestion points: Data is pulled via vitally:get_account_notes and vitally:get_account_conversations in SKILL.md (Step 1d).
  • Boundary markers: The instructions do not specify any delimiters or warnings to the agent to ignore instructions that might be embedded within the customer notes.
  • Capability inventory: The agent has access to query-run for SQL execution and multiple Vitally API tools.
  • Sanitization: There is no evidence of sanitization, filtering, or validation of the content retrieved from external records before it is interpolated into the synthesis workflow.
  • [DYNAMIC_EXECUTION]: The skill dynamically generates SQL queries for the query-run tool by interpolating variables (e.g., {externalId}) retrieved from Vitally API calls. While this is consistent with the skill's primary purpose, it creates a surface for injection if the source data in the CRM is compromised.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 12:47 AM
Security Audit — agent-trust-hub — account-handover