AgentPod is building a private, secure device for your AI agent.The AgentPod device is coming.Coming soonBe first
We checked this and rejected itsecurity

Document Skills: why it is not safe to use

We did not find clear security issues, but the skill still merits caution because our review only covered the published materials we scanned.

What we found

Our review of the published Document Skills materials did not identify a specific failing control. In the files we scanned, we found no hidden activation rules, no hardcoded credentials, no destructive commands, no privilege escalation behavior, and no signs of obfuscated or remotely fetched code. The materials also did not show access to unrelated credentials such as environment variables, keychains, browser data, or SSH keys. For document handling, the reviewed examples described local processing and writing results to user-specified output files, and we did not find evidence that document contents were sent to undisclosed third parties.

Even with those positive findings, this skill did not clear review as a full pass because our assessment is limited to the repository materials and examples we examined. That means the warning reflects review scope and operational caution, not a confirmed harmful behavior.

What to do instead

If you use this skill, treat it as suitable only for low-risk document tasks until you verify its behavior in your own environment. Test with non-sensitive files first, confirm what files it reads and writes, and restrict access to only the documents needed for the task.

Want the same outcome, safely? Use our checked skill instead.

Source: https://github.com/anthropics/skills

We report what our security review found at the time we checked, with the goal of keeping people safe. Projects change; if a maintainer has since fixed this, we are glad to recheck it. Email hello@agentpod.com.

Copied to clipboard. Paste it into your AI (ChatGPT, Claude, or your agent) to add the skill.