Study found production AI agents vulnerable to tool-chain and delegation failures
Review RAT-134 · Catalogue RAT-134
GovInfoSecurity reports aggregate susceptibility testing of 847 deployments and tool-chain vulnerabilities. It identifies no bounded live rogue incident, affected organization or established unauthorized impact; vulnerability rates are not incident counts.
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
- 2026-10-30T13:13:39.723255Z UTC
- Discovery channel
- User-supplied comparison audit
- Discovery source
- https://rogueaitracker.com/incidents/agent-toolchain-delegation-failure-study/
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.govinfosecurity.comReporting / interview · www.govinfosecurity.com
Decision history
- · Case-by-case comparison audit
RAT-134: out-of-scope. GovInfoSecurity reports aggregate susceptibility testing of 847 deployments and tool-chain vulnerabilities. It identifies no bounded live rogue incident, affected organization or established unauthorized impact; vulnerability rates are not incident counts. Source checks: read: https://www.govinfosecurity.com/autonomous-ai-agents-fail-in-real-world-deployments-a-31633. GovInfoSecurity reports aggregate susceptibility testing of 847 deployments and tool-chain vulnerabilities. It identifies no bounded live rogue incident, affected organization or established unauthorized impact; vulnerability rates are not incident counts.
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.