ticket-triage
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data (support tickets) which could contain malicious instructions designed to manipulate the agent's behavior.
- Ingestion points: Ingestion of raw support tickets occurs in
SKILL.mdunder the 'Workflow' section (Step 1: Receive and parse the ticket). - Boundary markers: The instructions do not specify the use of delimiters (like XML tags or triple quotes) or provide 'ignore embedded instructions' prompts to the agent when processing ticket content.
- Capability inventory: While the skill itself primarily describes text processing and classification logic, the agent executing these instructions may have capabilities to route data to different teams or generate responses that interact with customers.
- Sanitization: The skill mentions 'normalizing the text by stripping signatures, quoted replies, and boilerplate', but it lacks sanitization against natural language instructions embedded within the ticket body.
Audit Metadata