DeepSeek-driven agent tried to turn lab server into proxy node
Review RAT-121 · Catalogue RAT-121
Jesta’s original FAQ explicitly says a human operator weaponized DeepSeek to run the proxyjacking campaign. The lab observed real attack traffic and induced model self-identification; autonomous command execution does not make the operation independent of malicious human intent. Do not treat the headline’s AI-not-person framing as evidence of a rogue benign-task agent.
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
- 2026-10-30T13:13:38.965991Z UTC
- Discovery channel
- User-supplied comparison audit
- Discovery source
- https://rogueaitracker.com/incidents/jesta-darkreasoning-deepseek-proxyjacking/
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
- jesta.aiOriginal account · jesta.ai
Decision history
- · Case-by-case comparison audit
RAT-121: out-of-scope. Jesta’s original FAQ explicitly says a human operator weaponized DeepSeek to run the proxyjacking campaign. The lab observed real attack traffic and induced model self-identification; autonomous command execution does not make the operation independent of malicious human intent. Do not treat the headline’s AI-not-person framing as evidence of a rogue benign-task agent. Source checks: read: https://jesta.ai/blog/darkreasoning. Jesta’s original FAQ explicitly says a human operator weaponized DeepSeek to run the proxyjacking campaign. The lab observed real attack traffic and induced model self-identification; autonomous command execution does not make the operation independent of malicious human intent. Do not treat the headline’s AI-not-person framing as evidence of a rogue benign-task agent.
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.