The Agent Skills Directory

[SAFE]: The skill is primarily instructional, providing frameworks and best practices for conversion rate optimization. It does not request access to sensitive credentials, perform network operations to untrusted domains, or execute shell commands.
[PROMPT_INJECTION]: The skill includes instructions to read .agents/product-marketing-context.md to gather product context. While this is a surface for indirect prompt injection, the risk is negligible as the skill lacks tools or capabilities that could be exploited by malicious content within that file.
Ingestion points: Reads from .agents/product-marketing-context.md or .claude/product-marketing-context.md in SKILL.md.
Boundary markers: None; the agent is directed to read the file content directly.
Capability inventory: No risky capabilities such as network access, file writing, or shell execution are utilized.
Sanitization: No input validation or sanitization is performed on the ingested context file.

onboarding-cro