Aardvark finds real vulnerabilities with a continuously running GPT-5 security agent
OpenAI introduced Aardvark, a GPT-5 security agent that continuously analyzes repositories, validates vulnerabilities, proposes patches, and had already produced CVEs.
0 comments · 0 votes
Sign in to join the discussion →
No comments yet. Start the discussion.
Why it moved the index
Months of operation, validated exploitable flaws, sandbox testing, proposed patches, and recognized CVEs demonstrate practical AI-enabled cyber resilience.
Assessment history
-
R1
Away 39 · confidence 88
New source-verified practical cyber-resilience deployment absent from durable context.
11 Aug 2026
Share this page
-
DoomBench assesses “Aardvark finds real vulnerabilities with a continuously running GPT-5 security agent” as evidence moving away from doom, with magnitude 39 and confidence 88 out of 100 in the human resilience category.
-
The DoomBench assessment of “Aardvark finds real vulnerabilities with a continuously running GPT-5 security agent” is based on reporting from OpenAI and records the editorial rationale, source quality, attribution, and revision history.
-
DoomBench summarizes “Aardvark finds real vulnerabilities with a continuously running GPT-5 security agent” as follows: OpenAI introduced Aardvark, a GPT-5 security agent that continuously analyzes repositories, validates...
https://www.doombench.com/news/aardvark-finds-real-vulnerabilities-with-a-continuously-running-gpt-5-security-agent-2025-10-30