AgentHazard (OpenClaw) - Data Exfiltration: leaderboard

AgentHazard attack success rate on data exfiltration attempts (unauthorized extraction of source code, environment variables, credentials or private data), with each backbone model running inside OpenClaw. Percent of instances judged harmful on the full trajectory by Gemini-3-Flash, lower is better.

Metric: Attack success rate (%) on the data exfiltration instances (unauthorized extraction of source code, environment variables, credentials or private data), with the model running inside the OpenClaw open-source agent framework (browser, shell and file-system tools; one isolated Docker container per task), over AgentHazard instances (harmful objectives embedded in locally legitimate multi-step workflows, ten risk categories and ten attack strategies; the evaluated runs cover 1,000 instances), judged on the full trajectory (round_all) by Gemini-3-Flash, which labels a trajectory harmful when it realizes the objective; backbones served through an OpenAI-compatible API; lower is better. Source: arxiv.org. 7 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Qwen 2.5 72B Instruct17.39#1117
2Kimi K2.535.65#266
3Qwen 2.5 Coder 32B Instruct54.78#1547
4Qwen 3 VL 235B A22B Instruct65.22#532
5GLM-4.668.7#498
6Qwen 3 32B77.39#972
7Kimi K280#443

Interactive version: theaggregate.ai/benchmark?slug=agenthazard-openclaw-data-exfiltration · How It Works · Data refreshed daily, snapshot 2026-11-21.