AI Risk Atlas Prototype/DemoUnofficial independent experiment. Not an official xAI product. Scores can be wrong.

Back to signals
critical83% confidenceseed

Agent-led exploit in a live repo after a codegen assistant missed it.

CapabilityDomain knowledgeAffordanceImpact domainCap-adjacent
Quoted from XOECD AI Incidents Monitor@OECD_AIM17 Aug 2026, 12:00

Quoted text

An AI agent exploited a vulnerability missed by GitHub Copilot in a Snowflake repository, enabling unauthorized data access. AIM incident.

Read and engage with the original on X. This desk is not a republication feed.

Analyst rationale

Agent + codegen + data access. Maps to CVE, creds, injection.

Related signals