What changed in AI security, Dec 8 to Dec 14, 2025
Dec 8 to Dec 14, 2025 (ISO week 2025-W50). Weeks run Monday to Sunday in UTC.
29 records published, +7 on the previous week: 11 vulnerabilities (-1), 0 incidents (no change), 18 research items (+12), 0 news items (-4), 0 policy items (no change).
Critical and high advisories
Vulnerability records rated critical or high, newest first.- High
CVE-2025-67818: Weaviate OSS path traversal during backup restore
CVE-2025-67818NVD/CVE Database - Critical
CVE-2025-67511: Cybersecurity AI command injection through run_ssh_command_with_credentials
CVE-2025-67511NVD/CVE Database - Critical
CVE-2025-67510: Neuron MySQLWriteTool executes arbitrary SQL provided by the caller
CVE-2025-67510NVD/CVE Database - High
CVE-2025-67509: Neuron MySQLSelectTool read-only bypass allows file writes via INTO OUTFILE
CVE-2025-67509NVD/CVE Database - High
CVE-2025-33213: NVIDIA Merlin Transformers4Rec deserialization flaw in Trainer
CVE-2025-33213NVD/CVE Database - High
CVE-2025-64671: Copilot command injection allows local code execution by unauthorized attacker
CVE-2025-64671NVD/CVE Database
Exploitation signals
Vulnerabilities published in the week that are listed in the CISA Known Exploited Vulnerabilities catalog or have an EPSS score of 10% or more.No vulnerability published in this week is listed as exploited or has an EPSS score of 10% or more.
Packages that began delegating to a language model
Exposure Registry packages whose first release declaring an LLM SDK, agent framework or MCP dependency was published in the week.| Package | Ecosystem | LLM SDKs | Release | Released |
|---|---|---|---|---|
| openreward | PyPI | Anthropic SDK, Google Gemini SDK, OpenAI SDK | 0.1.0 |
Topics that moved
Largest increases over the mean of the 4 previous weeks, for topics with at least 3 records in the week.| Topic | Records | Weekly mean, previous 4 | Difference |
|---|---|---|---|
| AI agents | 5 | 1.3 | +3.8 |
| Adversarial machine learning | 4 | 0.3 | +3.8 |
Research
Peer-reviewed first, then newest. Showing 8 of 18.Exploring the Agentic Metaverse’s Potential for Transforming Cybersecurity Workforce Development
Peer-reviewedAIS eLibrary (Journal of AIS, CAIS, etc.)PPFPL: Cross-Silo Privacy-Preserving Federated Prototype Learning Against Data Poisoning Attacks
Peer-reviewedIEEE Xplore (Security & AI Journals)Learning Generalizable Representations for Deepfake Detection With Realistic Sample Generation and Dual Augmentation
Peer-reviewedIEEE Xplore (Security & AI Journals)M&M: Secure Two-Party Machine Learning Through Modulus Conversion and Mixed-Mode Protocols
Peer-reviewedIEEE Xplore (Security & AI Journals)Why Not Diversify Triggers? APK-Specific Backdoor Attack Against Android Malware Detection
Peer-reviewedIEEE Xplore (Security & AI Journals)Toward Understanding the Tradeoff Between Privacy Preservation and Byzantine-Robustness in Decentralized Learning
Peer-reviewedIEEE Xplore (Security & AI Journals)Security Analysis of WiFi-Based Sensing Systems: Threats From Perturbation Attacks
Peer-reviewedIEEE Xplore (Security & AI Journals)Blockchain-Enhanced Verifiable Secure Inference for Regulatable Privacy-Preserving Transactions
Peer-reviewedIEEE Xplore (Security & AI Journals)
Policy and regulation
Newest first.No regulatory or policy records were published in this week.
Generated from the AI Sec Watch database at . Every item links to its record.