case-auto-resolve

Pass

Audited by Gen Agent Trust Hub on Aug 3, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted data from service desk incidents to perform automated actions.
  • Ingestion points: The skill retrieves ticket content (short descriptions and full descriptions) from the incident table via the SN-NL-Search tool and REST API GET endpoints.
  • Boundary markers: The instructions do not define boundary markers (such as triple quotes or XML tags) or specific guidance for the agent to ignore instructions embedded within the ticket text.
  • Capability inventory: The agent has the capability to modify records (SN-Update-Record) and add work notes (SN-Add-Work-Notes), which could be misused if an attacker successfully influences the agent's logic via a malicious incident description.
  • Sanitization: While the skill implements a logic-based confidence scoring system, it lacks explicit sanitization or filtering to detect or neutralize prompt injection attempts within the input fields.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 3, 2026, 02:24 PM
Security Audit — agent-trust-hub — case-auto-resolve