ChatGPT Data Leakage via a Hidden Outbound Channel in the Code Execution Runtime

2026-03-30T20:52:25Z81f716ada7aaf26ab7d34ec195f88ec61fb6dfff289a16154e3e0bc6c4d1fdd5
AI-assisted-malwareChatGPTClaude CodeHandala HackSilver DragonVoidLinkai-assistant-securityapi-token-theftcheck-point-researchchinese-aligned-actorscode-execution-runtimedata-exfiltrationhidden-outbound-channeliranian-actorsremote-code-executionthreat-intel

What happened

Check Point Research published multiple intelligence pieces (Mar 2026) highlighting acute AI-security risks and recent threat activity. Key findings: researchers discovered a hidden outbound channel in the ChatGPT code-execution runtime that can leak user data and sensitive artifacts uploaded to assistants; Anthropic’s Claude Code contains critical configuration-based vulnerabilities enabling remote code execution and API-token exfiltration (CVE-2025-59536, CVE-2026-21852); AI-assisted malware development has matured (VoidLink framework) and AI services are being abused as covert C2/evasion. O

Why it matters

A reviewed impact interpretation has not been published for this record.

Evidence and limitations

Source ID
checkpoint_research
Record identifier
81f716ada7aaf26ab7d34ec195f88ec61fb6dfff289a16154e3e0bc6c4d1fdd5
Enrichment time
2026-03-30T20:52:25Z
AI-assisted enrichment
Yes

This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.