SafeClawBench

What is SafeClawBench?

SafeClawBench is a staged benchmark for tool-using LLM agent security with 600 adversarial tasks across six attack families, reporting three endpoints…

Released

2026-06-16

Evaluates

Agents, General AI, Language & Knowledge

Openness

unknown

Importer

Claire Radar

Review status

unreviewed

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2606.18356

SafeClawBench paper and dataset

• Paper https://arxiv.org/abs/2606.18356
• Dataset https://huggingface.co/datasets/sairights/safeclawbench

Which sources cite SafeClawBench?

• SafeClawBench

Related Agents, General AI, Language & Knowledge benchmarks

• $\tau^\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction
• 3DCodeBench
• 5-Dialects-BN
• A Large-Scale Dataset and Benchmark
• A3S-Bench
• AARRI-Bench
• ABC-Bench
• ActBench