Add Awesome-Embodied-AI-Safety (400+ papers on embodied AI agent safety)

[x-zheng16]

Hi,

We maintain Awesome-Embodied-AI-Safety, a survey repository with 400+ papers on safety risks in embodied AI agents (robots, autonomous vehicles, physical manipulators), organized across five safety layers: perception, cognition, planning, interaction, and agentic system safety.

As AI agents become more capable and autonomous, safety is a critical concern -- from adversarial attacks on agent perception and planning, to jailbreaking language-conditioned agents, to backdoor attacks on agent policies. Our repo provides a comprehensive catalog of these risks and defenses specifically for embodied/physical AI agents.

Would you consider adding it to your list?

• Awesome-Embodied-AI-Safety: 400+ papers on embodied AI agent safety -- adversarial attacks, jailbreaks, backdoors, and defenses across perception, cognition, planning, interaction, and agentic system layers.

Thank you for the excellent curation!