The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to automatically load context from external files located in the user's repository, which represents a vulnerability surface for indirect prompt injection.
Ingestion points: The agent is directed to read from .agents/product-marketing.md, .claude/product-marketing.md, or product-marketing-context.md before proceeding with the user's task.
Boundary markers: There are no instructions to the agent to treat this ingested content as data or to ignore any instructions found within those files.
Capability inventory: While the skill body is descriptive, the agent environment provides capabilities that could be exploited by malicious content in the ingested files.
Sanitization: The skill does not implement or recommend any sanitization or validation of the external context files.

free-tools