# Meta — How We Built Safety Into Muse

- Company: Meta (meta.com)
- Announced: 2026-09-08
- Category: not stated
- Coverage: not counted
- Announcement: no
- Group: routine
- Source: https://research.meta.ai/blog/security-and-safety-for-ai-agents-our-approach-with-muse
- Record: https://forck.live/items/9303-how-we-built-safety-into-muse
- Subject: Muse

Meta launched Muse, a personal agent that operates in a dedicated cloud virtual machine, with safety systems including isolated runtime cells, a Sentinel permission authority, and a bug bounty program offering up to $300,000.

## Evidence

Verbatim from https://research.meta.ai/blog/security-and-safety-for-ai-agents-our-approach-with-muse:

> Today we launched Muse — our personal agent.

---

Record: https://forck.live/items/9303-how-we-built-safety-into-muse
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md