n8n-validation-expert
Pass
Audited by Gen Agent Trust Hub on Mar 8, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface as it is designed to interpret and fix validation errors from data provided in workflow configurations.\n
- Ingestion points: Workflow and node configurations are ingested via functions like validate_node and validate_workflow (SKILL.md).\n
- Boundary markers: There are no instructions for the agent to use specific delimiters or to disregard instructions embedded within the JSON data being validated.\n
- Capability inventory: The skill guides the agent in using tools with write access to the user's environment, such as n8n_create_workflow and n8n_update_partial_workflow (SKILL.md).\n
- Sanitization: The skill does not mention sanitizing or escaping the content of node properties (e.g., descriptions or expressions) before processing them or feeding them to other tools.
Audit Metadata