duplicate-detection

Pass

Audited by Gen Agent Trust Hub on May 20, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: Vulnerability to indirect prompt injection is present because the skill retrieves and analyzes content from the text field in kb_knowledge and the comments field in kb_feedback. Malicious instructions embedded in these fields by an attacker could potentially influence the AI agent's analysis or actions.
  • Ingestion points: The skill reads article content from the text field of the kb_knowledge table and user feedback from the comments field of the kb_feedback table (Step 3 and Step 4).
  • Boundary markers: None identified in the provided procedure or examples to delimit untrusted article data from agent instructions.
  • Capability inventory: The skill includes the ability to update and retire ServiceNow records via the SN-Update-Record tool (Step 6).
  • Sanitization: No sanitization, filtering, or validation of the ingested article content is specified before analysis.
Audit Metadata
Risk Level
SAFE
Analyzed
May 20, 2026, 01:35 PM
Security Audit — agent-trust-hub — duplicate-detection