OpenClaw reveals meaty personal information after simple cracks

OpenClaw, an open-source AI agent platform, is reported vulnerable to indirect prompt injection that can be used to create attacker-controlled integrations and backdoors; researchers also found hundreds of ClawHub skills that leak credentials and dozens of malicious payloads capable of credential theft, backdoor installation, data exfiltration, and C2 deployment (including a demonstrated Sliver beacon proof-of-concept).