AI Risk Wire
Incidents
Risks
Controls
AI Risk Trends
Regulation
Newsletter
Contact us
5 controls that failed in an incident
Read for controls: 27 of 166 incidents.
Control
Who owns it
Times it failed
Most often failed under
Latest incident
Human approval for high-impact actions
Deployer
8
Hidden instructions hijack the agent
(
4
)
2026-09-29
Meta Muse agent disclosed Marketplace seller's home address and impersonated him to a buyer
Live behaviour monitoring
Deployer
4
Deception, concealment or sabotage
(
2
)
2026-08-04
AISI cyber evaluation: Mythos 5 agent exposed locally-hosted attack tooling to the public internet via tunnelling services
Least-privilege access for the agent
Deployer
1
Supply-chain compromise
(
1
)
2025-08-26
Nx s1ngularity: malicious npm packages drove local AI coding CLIs to enumerate secrets for exfiltration
Production kept apart from test
Deployer
1
Agent takes destructive or unapproved actions
(
1
)
2026-04-25
Cursor coding agent deletes PocketOS production database and backups
Output filtering and redaction
Deployer
1
Hidden instructions hijack the agent
(
1
)
2026-09-24
SalesBleed: 0-click data exfiltration from Salesforce Agentforce via indirect prompt injection in Web-to-Lead form