Everything Threat Times has reported about GitHub: 21 stories, newest first. Part of our AI coverage.
Researchers found that npm’s Trusted Publishing feature allowed attackers to publish malicious updates without stealing credentials by claiming abandoned repository names.
North Korea-linked group targets Ukrainian policy experts with malware delivered via trusted code hosting platforms.
The open-source mapping software project is enforcing a strict no-AI policy to protect maintainers from the overwhelming volume of low-quality pull requests.
GitHub explains how developers can use custom instructions, agents, and MCP servers to align AI assistance with project standards.
Researchers identify a state-aligned Iranian group using fake recruitment portals and custom malware to compromise aviation and telecom sectors in the Middle East.
AI firm dumps hundreds of preliminary math papers online, deepening rift with academic community over misaligned goals.
Threat actors used stolen maintainer credentials to inject malicious workflows into hundreds of repositories, stealing secrets and cloud keys.
Researchers link a fake video conferencing tool and abused GitHub repositories to a cyberespionage campaign targeting government entities in Israel and the Kurdistan Region of Iraq.
A new campaign by GhostAction has expanded to over 500 compromised accounts, injecting malicious workflows into tens of thousands of repositories since October 2026.
The self-hosted Talorys project, which runs on Cloudflare’s free tier, attracted significant attention and discussion on the Hacker News platform this week.
Socket research reveals a new wave of the GhostAction malware campaign stealing cloud and AI credentials from hundreds of GitHub repositories.
Version 2.4.0 of the self-hosted task manager Vikunja resolves a high-severity flaw where plaintext tokens allowed full account takeover if database access was compromised.
Researchers reveal a campaign using stolen maintainer credentials to inject malicious GitHub Actions into hundreds of open-source repositories.
GitHub has integrated a new AI-based classifier with Microsoft to detect unstructured secrets and block them before they enter repository history.
Black Lotus Labs reports a cryptomining botnet using a GitHub poem to hide command-and-control addresses, compromising over 3,400 servers.
New findings from GitHub indicate that hackathons remain the most effective environment for individuals to learn software construction in an era of reduced barriers.
A critical flaw in Praisonai’s AICoder component enables attackers to write files anywhere on the filesystem and execute commands via prompt injection.
The FakeGit malware campaign has returned, using over 17,600 fake GitHub repositories to distribute SmartLoader and the StealC infostealer to unsuspecting developers.
Palo Alto Networks researchers identify an Iran-nexus operation using fake Dubai Airport recruitment lures and GitHub-based command and control infrastructure to compromise systems.
OpenAI shares new findings on open mathematical problems from an internal frontier model, releasing code and research details publicly.
GitHub highlights a bug bounty researcher's approach to selecting features for security testing during Cybersecurity Awareness Month.