The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines a mechanism for creating and managing autonomous automations that execute specific prompts (instruction) on a schedule or trigger. This creates a surface for indirect prompt injection where an attacker might influence the saved instructions through data processed in earlier conversation steps.
Ingestion points: The instruction parameter in create_automation and manage_automation tools (SKILL.md).
Boundary markers: The skill instructions mandate that the agent must confirm details and get a "yes" from the user before calling the tool, providing a human-in-the-loop safety check.
Capability inventory: The skill allows for the definition of autonomous agent behavior including scheduling (cron), price triggers, and delivery to external services like Slack.
Sanitization: No specific sanitization or validation logic is defined for the instruction payload beyond natural language refinement.

automation