AI evaluation agents reached organizations they were not targeting
Pillar Security counted 21 public disclosures from July to mid-October involving OpenAI, Anthropic, Google and Meta agents.
64
Organization
Pillar Security
- Mentions · 7d1
- Mentions · 30d5
- All mentions10
- First seenSep 20, 2026
- Last mentioned5d ago
Timeline
10 articles · newest firstUnsloth loader can run Python even with trust disabled
CVE-2026-93348, scored 8.6, is fixed in Unsloth 2026.8.20 and Unsloth Zoo 2026.8.14.
62
OpenCode MCP sign-in flaw executes code on Windows and WSL
Builds 1.0.137 through 1.18.32 are affected; Anomaly fixed it in 1.18.33 without a CVE or advisory.
68
OpenCode 1.0.137–1.18.32 RCE silently fixed in 1.18.33
Pillar Security says a malicious MCP server can run code on Windows during OAuth, and scanners will not flag the flaw.
72
Microsoft and NVIDIA say open-weight AI shifts security burden to enterprises
Guidance covers 30 risk types, unsigned model files, and backdoors that behavioral tests may miss.
48
Unsloth’s model picker had a code-execution problem
Unsloth Studio executed remote Python when users merely selected a model, letting crafted repos run code on developer machines.
63
Unsloth Studio RCE Flaw Lets Malicious Hugging Face Models Execute Code
Patched Unsloth Studio flaw let malicious Hugging Face models execute Python code when selected, running with user privileges; users must upgrade to 2026.6.9.
50
Unsloth Fixes Arbitrary Code Execution Flaw Triggered by Malicious Hugging Face Models
Unsloth patched Studio 2026.6.9 after default remote-code trust let malicious Hugging Face models execute Python on inspection.
62
Growing the WIN AI Ecosystem with Agent Integrations
Wiz made its WIN partners MCP endpoint generally available so partner agents can pull live security context.
34
Researchers escape OpenAI Codex sandbox to run commands on host
Researchers found two OpenAI Codex sandbox escapes, including Heapjack, enabling unsandboxed command execution on developer machines; OpenAI fixed both within eight days.
68
Related CVEs
by mentions- Code injection RCE in Unsloth model loading
Appears with
- Hugging Faceorganization · 3
- Unslothvendor · 3
- Codex CLIproduct · 2
- Anthropicorganization · 2
- Googleorganization · 2
- Unsloth Studioproduct · 2
- Unslothorganization · 2
- OpenAIvendor · 2
- OpenCodeproduct · 2
- OpenAIorganization · 2
- Anomalyvendor · 2
- BleepingComputerorganization · 1
- NVIDIAvendor · 1
- International Energy Agencyorganization · 1
- Unsloth Zooproduct · 1
- Wiz MCPproduct · 1
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.