email-recommendation
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests untrusted data from external sources and uses it to generate email response recommendations for agents. A malicious customer could embed instructions in their email that the AI agent might follow during the recommendation generation process. \n
- Ingestion points: The skill retrieves external data from the
sys_email(Step 2) andsn_customerservice_case(Step 1) tables. \n - Boundary markers: The skill does not define explicit delimiters or 'ignore embedded instructions' prompts when processing the retrieved communication history. \n
- Capability inventory: The skill utilizes
SN-Query-Table,SN-Get-Record, and several ServiceNow REST API GET endpoints. \n - Sanitization: No sanitization, filtering, or validation steps for external content are specified before the data is used in response generation.
Audit Metadata