The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data in the form of marketing page content provided by the user. While this creates a potential surface for indirect prompt injection, the skill acts purely as an advisor generating text recommendations and does not invoke high-risk tools (such as shell commands or network exfiltration) that would allow an injection to escalate into a system-level compromise.
[DATA_EXPOSURE]: The instructions direct the agent to check for a local file at .claude/product-marketing-context.md to gain business context. This is a standard and safe practice for providing project-specific knowledge to an agent within a development environment, and no mechanisms for exfiltrating this data were found.

page-cro