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

Claude Skills Collection: why it is not safe to use

We found skills that can access authenticated Yahoo Fantasy sessions and private finance PDFs, so extra care is needed with sensitive data.

What we found

Our review did not find hidden instructions, hardcoded secrets, obfuscated content, privilege escalation, or undisclosed data transfers in the inspected repository materials. The documented skills mainly read user-provided files or user-requested web pages and write results locally.

The main concern is data sensitivity. At least one skill is designed to work through an authenticated Yahoo Fantasy browser session using Claude-in-Chrome. The collection also includes household finance workflows that process private financial PDFs. Based on the inspected docs and sampled files, this access appears tied to accounts and documents the user intentionally connects, not to hidden collection or export. Even so, these workflows can place sensitive personal or account data into the skill context, which raises privacy risk if used without careful scoping.

What to do instead

Use this collection only with accounts, browser sessions, and documents you are comfortable exposing to the skill context. Prefer test or low-sensitivity data where possible. Limit access to the minimum files and pages needed, review prompts and outputs for accidental inclusion of sensitive details, and avoid connecting high-risk financial or personal records unless that access is necessary.

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

Source: https://github.com/lyndonkl/claude

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.