AI agents compromised 440 PaperCut servers worldwide
Review RAT-098 · Catalogue RAT-098
GreyNoise’s original identifies a likely Russian-speaking malicious actor constructing a lab then launching hundreds of Codex-harness/DeepSeek agents at PaperCut victims. These are real compromises, but human-directed attacks. Failure of the attacker’s country exclusions does not establish an agent exceeding benign intent.
Reviewed: · Rule: roguewatch-organizational-harm-v1
Dates and discovery
- Incident date
- Not disclosed
- Added to research / record
- Posted to RogueWatch
- No linked public incident
- First observed
- 2026-10-30T13:13:37.798978Z UTC
- Discovery channel
- User-supplied comparison audit
- Discovery source
- https://rogueaitracker.com/incidents/papercut-ai-agent-global-intrusion/
Dates use America/Los_Angeles. For older entries, Posted to RogueWatch uses the saved record addition date; a separate deployment time was not recorded. Future posting timestamps record first public qualification. Source publication and incident occurrence are separate.
Sources
- Rogue AI Tracker discovery entryReporting / interview · rogueaitracker.com
- www.greynoise.ioOriginal account · www.greynoise.io
Decision history
- · Case-by-case comparison audit
RAT-098: out-of-scope. GreyNoise’s original identifies a likely Russian-speaking malicious actor constructing a lab then launching hundreds of Codex-harness/DeepSeek agents at PaperCut victims. These are real compromises, but human-directed attacks. Failure of the attacker’s country exclusions does not establish an agent exceeding benign intent. Source checks: read: https://www.greynoise.io/blog/ai-orchestrated-campaign-against-papercut-ng-mf. GreyNoise’s original identifies a likely Russian-speaking malicious actor constructing a lab then launching hundreds of Codex-harness/DeepSeek agents at PaperCut victims. These are real compromises, but human-directed attacks. Failure of the attacker’s country exclusions does not establish an agent exceeding benign intent.
Evidence source
A review decision applies only to the named scope. Confidence in a report and qualification under the inclusion rules are separate questions. Review records do not add incidents to the globe or leaderboard.