PPhying中文
Curated security research, vulnerabilities, advisories and tools for practitioners.

Hacking An AI Children's Toy: Remote Access to Every Conversation

josephthacker.com | blog | #ai-security | #data-exposure | #iot | #vulnerability-disclosure | #broken-access-control | #ai-toys | #child-safety
📄 Raw .md⬇️ Download .md🔗 Original ↗

Summary

Bondu's AI children's toy admin console was open to any Google login, exposing tens of thousands of child-toy conversation transcripts plus children's names, family details, and device data.
Published
Collected

Related coverage

blog · josephthacker.com
AI Models Are Not Safety-Tuned for Kids
AI safety tuning assumes adult users, yet children use AI apps daily without models knowing their age; Joseph Thacker argues age-appropriate response boundaries are an overlooked safety gap.
blog · portswigger.net
How Burp AT helped expose whistleblower reports via a critical vulnerability that was overlooked for years
A customer story on how Burp AT, Burp Suite's agentic testing capability, helped Orange Cyberdefense analyst Ray Huygen work through 66,000 lines of minified JavaScript to find a critical access flaw that exposed confidential whistleblower reports, and turn the finding into a confirmed exploit.
blog · blog.quarkslab.com
Overview of Passive Optical Networks (PONs) Security
Passive Optical Networks (PONs) connect end-users to infrastructure using optical fibre in the last kilometre (Fibre-to-the-x). This article provides a technical overview of the security features in ITU-T specifications: Gigabit-capable PON (GPON), 10-Gigabit-capable PON (XG-PON), 10-Gigabit-capable Symmetric PON (XGS-PON), Next-generation PON 2 (NG-PON2), and 50-Gigabit-capable PON (50G-PON). The analysis covers authentication schemes, key derivation, encryption methods, and associated security implications.
blog · bishopfox.com
Otto Support - Testing MCP Servers
Closing the otto-support series, the team uses nmap, a Nuclei template, and MCP Inspector to expose an authorization gap that lets any unprivileged user delete other users' tickets.
blog · depthfirst.com
Our Approach to Coordinated Vulnerability Disclosure
depthfirst's coordinated disclosure policy: every report is human-verified and includes a PoC and fix guidance; vendors get 30 days to patch, then a 7-day adoption window before publication.
blog · bishopfox.com
How to Engineer Secure Things: Past Mistakes and Future Advice
The recurring engineering mistakes behind insecure IoT products, starting with unvalidated firmware updates (see the TrackingPoint rifle), plus advice like root of trust and code signing.
Back