n8n-validation-expert
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill consists exclusively of markdown documentation files including SKILL.md, README.md, ERROR_CATALOG.md, FALSE_POSITIVES.md, and REVIEW_CHECKLIST.md. No executable scripts or hidden commands were found.
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to process n8n workflow configurations, which are untrusted external data sources. 1. Ingestion points: Workflow and node JSON objects passed to validation tools as defined in SKILL.md. 2. Boundary markers: The instructions lack explicit delimiters to separate untrusted workflow data from agent instructions. 3. Capability inventory: The skill leverages n8n_update_partial_workflow and n8n_autofix_workflow to modify workflow states. 4. Sanitization: No specific input sanitization or filtering is defined for the workflow data being analyzed.
- [SAFE]: Mentions of credentials and security risks in ERROR_CATALOG.md and FALSE_POSITIVES.md are instructional examples designed to help the agent detect and remediate these issues in user workflows, rather than representing actual leaks or vulnerabilities in the skill itself.
Audit Metadata