incident-triage
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 text data which presents an injection risk surface.
- Ingestion points: Untrusted ticket data (
short_description,description) is loaded into context via theSN-Query-Tabletool inSKILL.md. - Boundary markers: Absent. There are no delimiters or specific guidelines instructing the model to ignore potential prompt instructions embedded within the incident text fields.
- Capability inventory: The skill possesses read and write/update capabilities across incidents using tools like
SN-Query-Table,SN-Assign-Incident,SN-Update-Record, andSN-Add-Work-NotesinSKILL.md. - Sanitization: Absent. No filtering, validation, or escaping of the parsed incident content is performed prior to parsing or keyword execution.
Audit Metadata