n8n-validation-expert

Pass

Audited by Gen Agent Trust Hub on Feb 26, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection Surface.
  • Ingestion points: The skill processes potentially untrusted n8n node configurations and validation error messages provided by the user or external systems during the validation loop. (Files: SKILL.md, ERROR_CATALOG.md)
  • Boundary markers: There are no instructions defining the use of delimiters or security headers (e.g., 'ignore embedded instructions') when parsing external node data.
  • Capability inventory: The skill enables the agent to use workflow-modifying tools, specifically n8n_update_partial_workflow and n8n_autofix_workflow (SKILL.md), which could be manipulated if an injection is successful.
  • Sanitization: While the skill describes 'Auto-Sanitization' for the structural integrity of operators, it does not include sanitization or validation of the input content to prevent instruction-based prompt injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 26, 2026, 07:08 AM