LowResearchPreprintLLM-specific
Network-Exposed Configuration Attacks on LLM-Controlled Social Robots
- Published
- Record updated
Summary
Researchers show that the Pollen Robotics Reachy Mini Lite exposes a management interface that lets an unauthenticated attacker replace its personality settings and the cloud LLM's system prompt, without any adversarial conversational input. Eleven prompt-injection attempts were resisted, but both configuration-layer attack paths caused observable behavioural changes on physical hardware. The weakness maps to CWE-306 and CWE-1327 (CVSS v3.1 6.5, Medium), and a source review of 59 public repositories found missing authentication in 10 of 11 source-verifiable genuine applications.
Topics
Related items
- HighCVE-2026-77283: obsidian-sync-mcp HTTP /mcp endpoint DNS rebinding and cross-origin vault accessSimilar attack · NVD/CVE Database
- HighCVE-2026-75612: yt-mcp arbitrary file read through add_attachment file_pathSimilar attack · NVD/CVE Database
- MediumCVE-2026-75518: DOMShell MCP server eval tool bypasses write permission checksSimilar attack · NVD/CVE Database
- HighCVE-2026-62964: Gortex arbitrary file write through MCP file toolsSimilar attack · NVD/CVE Database
- HighCVE-2026-127222: SGLang missing authorization on POST /load_lora_adapter_from_tensorsSimilar attack · NVD/CVE Database