Default context-file reading can expose confidential marketing information to the AI context
Source references: 1The Skill instructs the agent to read product-marketing files from several conventional locations before asking the user, without first confirming what is needed for the current task. Such files may contain positioning, customer, pricing, or launch strategy information.
If accessible, their contents may enter model context, logs, or generated documents even when the RevOps question does not require all of them. The source does not show network exfiltration, but it broadens exposure of sensitive business material.
This is a live Skill instruction: if a named file exists, the AI reads it before asking questions. Its contents therefore enter the model context even when the user only needs narrow RevOps advice. The source does not require prior scope confirmation or exclusion of sensitive fields. Risk depends on what the file contains and the AI service's data handling; this does not prove disclosure occurred. Users can ask for opt-in access, task-limited excerpts, and clear retention boundaries.
**Check for product marketing context first:**If `.agents/product-marketing.md` exists (or `.claude/product-marketing.md`, or the legacy `product-marketing-context.md` filename, in older setups), read it before asking questions. Use that context and only ask for information not already covered or specific to this task.