n8n-workflows-master

Pass

Audited by Gen Agent Trust Hub on Feb 19, 2026

Risk Level: SAFE
Full Analysis
  • [Data Exposure & Exfiltration] (SAFE): The skill demonstrates strong credential hygiene. Both the provided sample workflow and the internal system prompts for the 'Build Captain' persona emphasize the use of named credential placeholders (e.g., 'CRED_API_KEY_HEADER') and explicitly forbid hardcoding actual secrets or API keys within the generated JSON exports.
  • [Indirect Prompt Injection] (SAFE): The skill processes untrusted user specifications to generate executable-like JSON files, creating an ingestion surface. However, it incorporates a logical mitigation layer via the 'Security Architect' persona.
  • Ingestion points: Plain language user specifications for workflow creation (SKILL.md).
  • Boundary markers: Role-based instruction sets provide functional separation between creation and security review tasks.
  • Capability inventory: File writing to the './workflows/' directory and generation of n8n node logic.
  • Sanitization: The 'Security Architect' persona is explicitly tasked with auditing the final JSON for risks such as SSRF, PII exposure, and insecure authentication before 'shipping'.
  • [Dynamic Execution] (SAFE): While n8n workflows can execute custom scripts via 'Code' nodes, the skill's instructions mandate prioritizing deterministic, built-in nodes and require a security review of any generated logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 19, 2026, 06:26 AM