PuzzleMask: Abusing Plain Prose as a Covert AI Attack Vector
Check Point details PuzzleMask, a plain-prose technique that bypasses LLM gatekeeper policy checks, letting hidden payloads reach target models unreviewed.
55
Product
GPT-OSS-Safeguard 20b
- Mentions · 7d0
- Mentions · 30d0
- All mentions1
- First seenSep 10, 2026
- Last mentionedSep 10, 2026
Timeline
1 articles · newest firstAppears with
- Llama Guard 3ai model · 1
- GPT-4o-miniai model · 1
- Anthropicorganization · 1
- GPT-OSS-Safeguard 20bai model · 1
- Check Point Researchvendor · 1
- Claude 3 Haikuai model · 1
- Metaorganization · 1
- GPT-5product · 1
- Claude 3 Haikuproduct · 1
- GPT-4o-miniproduct · 1
- GPT-5ai model · 1
- Llama Guard 3product · 1
- OpenAIorganization · 1
- Check Point Researchorganization · 1
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.