# Has an AI agent ever broken a rule at your company that nobody wrote down? — retained record

> Retained reddit record from reddit:AI_Agents.

Canonical: https://highsignal.app/data/records/3d1be74ee1db9551

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


source: [reddit](https://highsignal.app/data/reddit) (reddit:AI_Agents)  published 2026-11-03 04:47 ingested 2026-11-04 02:36

[original source ↗](https://www.reddit.com/r/AI_Agents/comments/1xmju2r/has_an_ai_agent_ever_broken_a_rule_at_your/)

## retained text



```
I’m a high school student working on a startup around AI agents and production changes. Not selling anything and there’s no link, I’m just trying to learn from people who’ve actually dealt with this.

I keep reading stories like the Replit agent that deleted a database during a code freeze, or the Cursor agent that wiped a company’s database and backups while it was supposed to be working on staging. In a lot of these, nothing actually blocked the agent. It just didn’t know something everyone on the team knew.

Two questions:

1. Has an agent (or any automation) ever done something in prod that was technically allowed but obviously wrong to anyone who’d been there a while? What did it miss?
2. If you do let agents make some changes on their own, how did you decide which ones? What had to be true first?

Even short answers help a lot. And if you’ve got a good story and wouldn’t mind a 15 min call, I’d really appreciate it.
```



document fetched 2026-11-04 00:19 · canonical https://reddit.com/r/AI_Agents/comments/1xmju2r/has_an_ai_agent_ever_broken_a_rule_at_your
