n8n-validation-expert
Pass
Audited by Gen Agent Trust Hub on Oct 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes output from validation tools which may contain untrusted data if the workflow being analyzed is controlled by an attacker.
- Ingestion points: External data enters the context through the results of the
validate_nodeandvalidate_workflowtool calls, specifically within error messages and warnings found inSKILL.mdandERROR_CATALOG.mdinstructions. - Boundary markers: The instructions do not mandate the use of delimiters (e.g., XML tags or backticks) to wrap validation output, which increases the risk that embedded instructions in error messages could be interpreted as agent commands.
- Capability inventory: The skill possesses significant modification capabilities, including
n8n_update_partial_workflow,n8n_autofix_workflow, andn8n_manage_credentialsacross the workflow configuration files. - Sanitization: There are no instructions provided to sanitize, escape, or filter external content from validation responses before the agent interpolates this data into its reasoning process.
Audit Metadata