Public research record · Editorial guide updated 2026-10-29

Multi-agent framework harvested thousands of credentials in six hours

Review RAT-099 · Catalogue RAT-099

RogueWatchOut of scope

GTIG’s original threat-intelligence account describes adversaries adopting multi-agent scanning, credential harvesting, and cloud abuse. This is attacker-directed agentic tradecraft; no specific agent acted against the directing operator’s authorization.

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
2026-10-30T13:13:37.542263Z UTC
Discovery channel
User-supplied comparison audit
Discovery source
https://rogueaitracker.com/incidents/gtig-multi-agent-credential-harvesting/

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

Decision history

  1. · Case-by-case comparison audit

    RAT-099: out-of-scope. GTIG’s original threat-intelligence account describes adversaries adopting multi-agent scanning, credential harvesting, and cloud abuse. This is attacker-directed agentic tradecraft; no specific agent acted against the directing operator’s authorization. Source checks: read: https://cloud.google.com/blog/topics/threat-intelligence/from-prompting-to-autonomy-the-evolution-of-adversarial-ai. GTIG’s original threat-intelligence account describes adversaries adopting multi-agent scanning, credential harvesting, and cloud abuse. This is attacker-directed agentic tradecraft; no specific agent acted against the directing operator’s authorization.

    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.