In the news
CWE-bench v2: More coverage, quality, and transparency
Hugging Face · Published · 3 min read
In 30 seconds
- What happened
- CWE-bench v2 released with 100 public cybersecurity tasks covering 91 vulnerability types for evaluating AI agent capabilities.
- Why it matters
- Security engineers and AI researchers building or testing autonomous agents that find and fix real vulnerabilities in code.
- Watch out
- Even top models fail 19% of tasks on every attempt; benchmark covers only known vulnerability classes, not novel attack surfaces.
- rag
The patterns behind this
Each one covers how the technique works, when it earns its cost, and where it breaks.
The Agent Architect
One pattern, one tradeoff, one production failure story. A short weekly briefing for people building agentic systems.
Weekly email, one-click unsubscribe. We only use your address to send the briefing.