catalog-item-generation
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted natural language input from users to generate service catalog items, variables, and choices, creating a potential surface for indirect prompt injection or schema confusion if malicious instructions are embedded within descriptions.
- Ingestion points: User-provided natural language descriptions parsed in Step 1 and processed dynamically in Examples 1, 2, and 3.
- Boundary markers: Absent; there are no specific delimiters or guardrail instructions directing the model to ignore embedded commands or instructions within the user input.
- Capability inventory: The skill leverages powerful capabilities including
SN-Create-Record,SN-Update-Record, andSN-Execute-Background-Script(which runs server-side JavaScript within the target environment). - Sanitization: Absent; the procedures do not specify input validation, escaping, or structural sanitization before interpolating user descriptions into record fields or script blocks.
Audit Metadata