The Agent Skills Directory

[SAFE]: The skill is composed entirely of natural language instructions and design guidelines. It does not contain any executable code, shell commands, or network-enabled tools.
[INDIRECT_PROMPT_INJECTION]: The skill instructions include a prompt to read project-specific context files (e.g., .agents/product-marketing-context.md). While this represents a data ingestion surface, the skill lacks any impactful capabilities—such as file writing, network access, or command execution—that could be leveraged if the ingested context contained malicious instructions.

form-cro