The Agent Skills Directory

Indirect Prompt Injection (SAFE): The skill instructs the agent to read context from a local file .claude/product-marketing-context.md if it exists. While this is a data ingestion point, the skill itself lacks any tools, script execution, or network capabilities that could be exploited by malicious content within that file. It is strictly advisory in nature.
Unverifiable Dependencies (SAFE): No external packages or remote scripts are referenced or installed by this skill.
Data Exposure (SAFE): The skill asks for business metrics (LTV, CAC, budget) to provide advice, but it does not contain code to exfiltrate this data or access sensitive system files.

referral-program