The Agent Skills Directory

Category 8: Indirect Prompt Injection (LOW): The skill interacts with the n8n REST API, creating an indirect prompt injection surface where malicious data in workflow names or execution outputs could influence agent behavior.
Ingestion points: scripts/n8n_api.py fetches data via list_workflows, get_workflow, and list_executions endpoints.
Boundary markers: Absent; the script returns raw JSON output directly to the agent without delimiters or safety headers.
Capability inventory: scripts/n8n_api.py can create, update, and delete workflows, as well as trigger executions with custom data payloads.
Sanitization: None; the tool assumes the API return data is trusted and does not sanitize strings before presenting them to the LLM.
Category 2: Data Exposure (SAFE): The skill correctly uses environment variables (N8N_API_KEY) for authentication and includes only placeholders in its documentation. No hardcoded credentials were detected.

n8n