AI frenzy feeds credential chaos, secrets leak through code, tools, and infrastructure
2026-03-27T08:52:10Z•b4d641f5b5d996ef47ac5bfe17397aa16925470c3119808afc33b8a46cedec90
AI chip smugglingAI training dataBPFDoorCopilotGitGuardianGitHubGoogleOpenAIPQCRed MenshenRedLineReddit bot verification','RSACTailsTor bridgescredential sprawldata privacyexport controlshardcoded secretsinfostealerpassword managerpost-quantum cryptographyquantum threatssafety bug bountysecretstelecom security
What happened
This feed highlights rising AI-era security risks and defensive moves: GitGuardian reports 28.65 million new hardcoded secrets exposed in public GitHub commits in 2025 and notes exposure is spreading into internal repos and infrastructure, exacerbating credential sprawl. Platform and policy shifts include GitHub’s change to use Copilot interaction data for model training (opt-out model) and OpenAI launching a public Safety Bug Bounty to surface model abuse. Operational security stories: Rapid7 released a BPFDoor detection script to find stealthy Red Menshen implants affecting telcos, a second/
Why it matters
A reviewed impact interpretation has not been published for this record.
Evidence and limitations
- Source ID
- helpnetsecurity
- Record identifier
- b4d641f5b5d996ef47ac5bfe17397aa16925470c3119808afc33b8a46cedec90
- Enrichment time
- 2026-03-27T08:52:10Z
- 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.