The Agent Skills Directory

[SAFE]: The skill content is purely instructional, focusing on best practices for form conversion rate optimization. It does not perform any dangerous shell commands, network operations, or sensitive data access.
[INDIRECT_PROMPT_INJECTION]: The skill instructions include reading from an external context file, which is a potential surface for indirect prompt injection.
Ingestion points: The agent is instructed to read .claude/product-marketing-context.md to gain context for the task.
Boundary markers: The skill does not provide instructions to the agent to treat the context file content as untrusted or to ignore embedded instructions.
Capability inventory: The skill has no tools, shell commands, or network capabilities defined, which eliminates the risk of an injection being exploited for harmful actions.
Sanitization: No sanitization is performed on the input from the context file before it is used by the agent.

form-cro