The Agent Skills Directory

[COMMAND_EXECUTION]: The skill references n8n nodes with high-privilege capabilities, including Execute Command for running shell commands and SSH for remote command execution as listed in references/n8n-nodes-masterlist.md.
[REMOTE_CODE_EXECUTION]: Documentation includes references to Code and Function nodes which allow for the execution of custom JavaScript and Python within workflows.
[EXTERNAL_DOWNLOADS]: The automation guide in references/n8n-workflow-automation-guide.md provides instructions for installing n8n via npm install and docker run, which involves downloading software from external registries.
[DATA_EXFILTRATION]: The skill describes workflows involving HTTP Request nodes and webhooks, which create an attack surface for data exfiltration if workflows are improperly configured with sensitive data.
[PROMPT_INJECTION]: The skill identifies a surface for indirect prompt injection within n8n workflows.
Ingestion points: Data enters the agent context via Webhook, HTTP Request, and various app-specific trigger nodes documented in references/n8n-nodes-masterlist.md.
Boundary markers: The prompt templates in references/n8n-ai-agent-prompt-formula.md lack explicit delimiters or instructions to ignore embedded commands within interpolated variables.
Capability inventory: High-privilege nodes like Execute Command, SSH, and Code are documented.
Sanitization: No explicit sanitization logic is demonstrated in the provided examples to handle untrusted input.

n8n-automation