The Agent Skills Directory

[Command Execution] (SAFE): The skill utilizes standard curl commands to interact with the n8n API. There are no signs of malicious command injection, privilege escalation, or unsafe pipe-to-shell patterns.
[Data Exposure & Exfiltration] (SAFE): No hardcoded secrets were detected. The skill correctly references environment variables (N8N_API_KEY) for authentication and does not attempt to access sensitive local system files.
[Prompt Injection] (LOW): The skill is susceptible to indirect prompt injection (Category 8) because it is designed to ingest and process untrusted external data.
Ingestion points: SKILL.md identifies 'Existing workflow JSON' and 'Sample payloads' as optional inputs for the design and operation tracks.
Boundary markers: Absent. The instructions do not provide delimiters or warnings to the agent to disregard instructions potentially embedded within the input JSON.
Capability inventory: The skill provides templates for executing curl requests that can modify workflows (/activate) or trigger external actions via webhooks.
Sanitization: Absent. No explicit validation or sanitization steps are defined for the incoming data payloads.

n8n-hub