The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to create persistent rule files (.mdc) in the .cursor/rules/ directory by inferring requirements from previous conversation context. This creates a surface where malicious instructions provided by a user or external data in the conversation history could be turned into persistent agent instructions.
Ingestion points: The skill explicitly directs the agent to 'Infer rules from what was discussed' in previous conversation context.
Boundary markers: The provided templates and instructions lack specific boundary markers or 'ignore' instructions to prevent the agent from adopting adversarial prompts embedded in the context.
Capability inventory: The skill utilizes the agent's ability to create and write markdown files to the local file system.
Sanitization: No sanitization or validation steps are defined to filter out potentially harmful instructions before they are committed to persistent rule files.

create-rule