The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest data from existing n8n workflows to support discovery and reuse, creating a surface for indirect prompt injection where malicious instructions embedded in workflow descriptions or node parameters could influence the agent.
Ingestion points: Data returned by the n8n_list_workflows and n8n_get_workflow tools (referenced in SKILL.md and NAMING_AND_DISCOVERY.md).
Boundary markers: No specific instructions are provided for the agent to use delimiters or ignore embedded instructions when processing the output of these tools.
Capability inventory: The skill utilizes tools with significant operational impact, including n8n_update_partial_workflow (which can modify any part of a workflow), activateWorkflow (which deploys changes), and n8n_manage_datatable (which can read/write external data state).
Sanitization: There are no explicit instructions for sanitizing or escaping content retrieved from external workflows before the agent interpolates it into new workflow definitions or its own reasoning process.
[DYNAMIC_EXECUTION]: The skill facilitates the creation and modification of executable logic within the n8n environment by generating structured JSON payloads for workflow nodes and expressions.
Evidence: The agent is instructed to use n8n_update_partial_workflow to programmatically build workflows by adding nodes and defining parameters, including complex JSON structures for workflowInputs (as seen in SKILL.md) and expressions (as seen in README.md). While this is the intended purpose of the skill, it represents the dynamic generation of execution logic from templates based on the agent's interpretation of requirements.

n8n-subworkflows