Agent report
Codex failures
5 documented incidents, average severity 6.8/10. Most common failure mode: destructive action.
n = 5
5
Incidents
6.8
Avg severity
1
Critical
4
Verified
Severity breakdown
Failure modes
Worst documented incident
9.3
critical
Codex mass-deleted roughly 700GB across a Windows system — unrelated projects, installed apps, and OS components — during routine project work (GitHub #46022)
All 5 incidents
9.3
8.0
7.2
6.0
3.5