# Am I the only one AFRAID of skills? — retained record

> Retained reddit record from reddit:ClaudeAI.

Canonical: https://highsignal.app/data/records/368da4d647d4276a

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:ClaudeAI)  published 2026-11-02 10:45 ingested 2026-11-03 02:37

[original source ↗](https://www.reddit.com/r/ClaudeAI/comments/1xltysk/am_i_the_only_one_afraid_of_skills/)

## retained text



```
On one hand, the new models are just great on their own without any extra skills.

Single shot prompts can already produce incredible videos and apps.

On the other hand, I want to use skills but I'm worried about context bloat. And then there's the security side of things, I keep hearing about malicious skills doing everything from prompt injection to stealing credentials and running wild.

I've used skills and gotten some great results, only to discover later that they're listed on [skills.sh](http://skills.sh) with security warnings.

So is it just me? Am I being overly cautious or is it time to get a little paranoid?

Do you use any kind of "antivirus" or security scanner for skills? How do you decide which skills are safe to use? Aren’t skills essentially external dependencies? If so, couldn’t newly discovered vulnerabilities turn a previously trusted skill into a security risk over time?
```



document fetched 2026-11-03 00:22 · canonical https://reddit.com/r/ClaudeAI/comments/1xltysk/am_i_the_only_one_afraid_of_skills
