The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest untrusted user input and pass it to internal tools for agent and workflow generation.
Ingestion points: The skill captures the user's request and forwards it verbatim to the build-agent tool via the message parameter as described in the "Faithful handoff" section.
Boundary markers: There are no specific instructions to use delimiters (such as XML tags or triple backticks) to isolate the user's raw input from the tool's execution context.
Capability inventory: The skill has access to powerful administrative tools including build-agent, build-workflow, and data-tables, which are capable of creating and modifying functional components of the n8n instance.
Sanitization: The instructions explicitly mandate forwarding the user's wording "as close to verbatim as possible" without providing a mechanism for filtering or sanitizing potentially malicious instructions embedded in those requests.
[SAFE]: The skill's operations are consistent with its primary purpose as a platform orchestrator. The tools and workflows mentioned are internal to the n8n ecosystem and follow established architectural patterns for agent building.

agent-builder