playbook-recommendations
Pass
Audited by Gen Agent Trust Hub on Aug 3, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted data from ServiceNow incident and case records, which creates a surface for indirect prompt injection attacks. Evidence: 1. Ingestion points: The skill fetches record context including short descriptions and symptoms using SN-Get-Record and SN-Query-Table (Steps 1, 2, and 8). 2. Boundary markers: There are no instructions to the agent to treat the retrieved record content as data or to use delimiters to prevent the execution of instructions embedded in those records. 3. Capability inventory: The skill has the capability to write back to the database using SN-Update-Record and lists Bash as a native tool in its configuration. 4. Sanitization: No sanitization, filtering, or validation logic is implemented for the external content before it is analyzed.
Audit Metadata