Agent report

Unknown agent failures

5 documented incidents, average severity 8.2/10. Most common failure mode: security vulnerability.
n = 5

Not a single vendor

These reports describe multi-agent or unattributed failures and are never counted toward any individual vendor.
5
Incidents
8.2
Avg severity
3
Critical
5
Source checked

Severity breakdown

Stacked severity breakdown per agent. Longer bar = more incidents.

Failure modes

Distribution of failure modes across all documented incidents.

Worst documented incident

10.0critical

An AI vendor's automated evaluation sandbox was prompt-injected into handing over its production API keys for multiple AI providers, which the attacker then used against the vendor and ~30 other AI companies (Anthropic GTG-50020)

All 5 incidents

10.0
An AI vendor's automated evaluation sandbox was prompt-injected into handing over its production API keys for multiple AI providers, which the attacker then used against the vendor and ~30 other AI companies (Anthropic GTG-50020)
10.0
Vibe-coded Tea app leaked 72,000 IDs and selfies plus 1.1M private messages from an unsecured bucket
9.0
Vibe-coded Moltbook exposed 1.5M API keys and 35,000 user emails via misconfigured database
7.0
AI vibe-coded Next.js app pinned vulnerable dependency — cryptominer compromised production server
5.0
An AI agent spun up duplicate CloudFormation stacks on every error and ran up a $6,531 AWS bill