The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is configured to use curl as its primary execution command and provides complex shell script examples in API_REFERENCE.md. These examples include loops and the use of jq to process API responses and automate actions across multiple workflows.
[DATA_EXFILTRATION]: The skill facilitates the transfer of data between the local environment and an external n8n host. While intended for management, this capability allows the agent to transmit the N8N_API_KEY and other sensitive environment variables to the configured $N8N_HOST.
[DATA_EXPOSURE]: The skill specifically includes endpoints for the n8n Credentials API. This allows the agent to list and retrieve details of credentials (such as API tokens for third-party services like Slack) stored within the n8n platform, representing a significant access level to the user's secrets management layer.
[INDIRECT_PROMPT_INJECTION]: The skill interacts with external data sources that may be attacker-controlled.
Ingestion points: Data enters the agent context via GET /api/v1/workflows and GET /api/v1/executions which fetch workflow names, node configurations, and execution logs from the n8n host.
Boundary markers: The instructions lack explicit delimiters or warnings to the agent to ignore instructions embedded within the retrieved data.
Capability inventory: The skill has the capability to execute shell commands (curl), delete resources on the n8n host, and trigger arbitrary workflows.
Sanitization: There is no evidence of sanitization or validation of the data retrieved from the API before it is processed by the agent.

n8n