The Agent Skills Directory

[SAFE]: The skill implements explicit privacy safeguards by forbidding the agent from asking for credentials, API keys, or passwords in chat.
[SAFE]: The skill restricts the agent's data gathering to user interactions, explicitly stating that it should not run commands or read files to find user information.
[INDIRECT_PROMPT_INJECTION]: The skill accepts free-text descriptions from users to generate workflow suggestions and configurations. While this is the intended purpose, it represents a surface where user-provided text could attempt to influence the agent's workflow generation tool.
Ingestion points: User replies to onboarding questions and free-text inputs in the ask-user cards in SKILL.md.
Boundary markers: The skill does not define specific delimiters for isolating user-provided task descriptions.
Capability inventory: The skill uses the build-workflow tool to create n8n workflows based on user input.
Sanitization: There is no instruction to sanitize or validate the user's input before workflow generation.
Mitigation: The skill includes a mandatory confirmation step requiring the user to explicitly approve the build action ("Yes, build it"), which serves as an important human-in-the-loop security check.

suggest-automations