Last updated: November 2, 2026 at 1:20 PM UTC
All 957 Vulnerability 399 Breach 144 Threat 407 Defense 7
Tag: github (24 articles)Clear

Gheist campaign uses stolen GitHub tokens to leak private repository secrets through Actions logs

Wiz detailed Gheist, an automated campaign that uses stolen GitHub tokens to push a sync-check.yml workflow to a short-lived ci-sync branch, then reads the job's secrets from the run log. The workflow base32-encodes every secret the job can access, defeating GitHub's log masking, downloads the log through the API, and deletes the run seconds later, so nothing leaves through an attacker server. Since September 15, Wiz recovered more than 13,500 planted branches across more than 850 compromised accounts, succeeding in private repositories and GitHub Enterprise, where GitHub's hold on suspicious runs does not apply. Rotate every reachable secret, since deleting branches does not undo exposure.

Check
Hunt for ci-sync branches containing only a sync-check.yml workflow and runs deleted shortly after finishing, and rotate every secret those repositories could access.
Affected
Private GitHub repositories and Enterprise servers reachable with a stolen token can have all their accessible secrets leaked through a planted workflow's run logs.
Fix
Rotate repository, organization, and environment secrets, revoke the tokens that pushed commits, prefer short-lived fine-grained tokens, and require reviewers on environments holding production secrets.

Attackers backdoor five OVHcloud CLI releases on GitHub to steal developer secrets from disk

OVHcloud disclosed that five releases of its open-source command-line tool, versions 0.11.0 through 0.15.0, were backdoored on GitHub to steal secrets from disk. An attacker who had controlled a developer's GitHub account for more than half a year modified the existing releases because GitHub release immutability was disabled, uploading locally built malware without touching the CI/CD pipeline, and planted a hidden preinstall.js loader. OpenSourceMalware linked it to the GlassWorm campaign. The malicious builds covered macOS, Linux, and Windows and were live from 14:47 to 16:21 UTC on October 19. Source builds and Docker images were unaffected, and users should revoke secrets.

Check
Identify machines that ran OVHcloud CLI versions 0.11.0 through 0.15.0, remove or update the binary, clean the systems, and revoke all secrets they could reach.
Affected
Developers who installed the backdoored OVHcloud CLI releases on macOS, Linux, or Windows had secrets stolen from disk by locally built malware uploaded to GitHub.
Fix
Update to a clean OVHcloud CLI release, rotate all exposed secrets, enable GitHub release immutability on your own repositories, and inspect hosts for the preinstall.js loader.

npm Trusted Publishing slug-reuse flaw exposed over 1,500 packages to unauthorized takeover

Bloom Security disclosed a flaw in npm's Trusted Publishing, which binds a package to a GitHub repository and workflow. npm recorded the binding by repository slug rather than GitHub's permanent numeric IDs, so once a bound repository or owner was renamed, transferred, or deleted, an attacker could claim the freed slug, add a matching workflow, and receive a publish credential, even forging a valid provenance attestation. Bloom counted more than 1,500 exposed packages with over 10 million total downloads. GitHub, notified in June through HackerOne, shipped a fix in October so unvalidated configurations expire after 48 hours. It is unknown whether anyone exploited the weakness in the wild.

Check
Run npm trust list to confirm each bound repository still belongs to you, rebind any pointing to renamed or transferred repositories, and remove bindings before deleting a repository.
Affected
npm packages bound through Trusted Publishing to a repository or owner that was later renamed, transferred, or deleted could be hijacked by an attacker claiming the freed slug.
Fix
Verify and rebind Trusted Publishing configurations, check that package provenance attestations name the expected repository, and apply the same review to GitLab-based Trusted Publishing.

A single GitHub issue could reach CI secrets across major AI coding agents

Novee Security showed at Black Hat that a GitHub issue opened by an account with no repository access could reach the CI runners behind major AI coding agents in their default configurations, tested against Claude Code, Gemini CLI, and Codex. The strongest, a Gemini CLI container-launcher command injection scored 10.0, runs code on the CI host before the sandbox starts. In Claude Code, a validator that stripped quoted text let a payload in a Git flag reach the runner, and a separate flaw leaked an API key through a download counter. Untrusted issue content reaching an agent that holds secrets and tools in the same runtime is the shared weakness.

Check
Update Gemini CLI to 0.39.1 and Claude Code to 2.1.163, and review any workflow where an AI agent runs automatically on issues or pull requests from untrusted users.
Affected
Teams running AI coding agents on public repositories in default configurations (CVE-2026-12537, CVE-2026-54316); an unprivileged GitHub issue or pull request can reach CI runners and expose workflow secrets and tokens.
Fix
Patch the agents, restrict their tools with allowlists rather than blocklists, give triage and review agents read-only tokens, and keep separate agent runs from sharing writable directories.

GitHub delays Dependabot version updates to keep poisoned packages out

GitHub is adding a default three-day cooldown before Dependabot opens pull requests for new package versions, aimed at supply chain attacks where a poisoned release spreads through automated updates before anyone catches it. Security updates that answer a known advisory still ship immediately; only routine version updates wait. GitHub points to the September 2025 compromise of chalk, debug, and other packages, whose crypto-stealing versions were live for roughly two hours, and notes its advisory database logged more than 6,500 npm malware advisories in the year to May 2026, around eighteen a day. Most malicious releases are caught within hours, so a short delay filters out the majority.

Check
If you use Dependabot, confirm the cooldown is enabled and consider tuning the window in dependabot.yml, and apply similar delay logic to any other automated dependency tooling you run.
Affected
Projects with automated dependency updates that pull new releases immediately; a poisoned version of a popular package can reach reviewers and installs during the short window before it is caught and yanked.
Fix
Adopt a cooldown on version updates while keeping security fixes immediate, pin and verify dependencies, review update pull requests rather than auto-merging, and monitor for advisories on packages you rely on.

FakeGit floods GitHub with fake AI skills and MCP servers that drop malware

Researchers at Island uncovered FakeGit, a campaign running roughly 7,600 malicious GitHub repositories from about 6,600 lookalike developer profiles, of which more than 800 pose as AI skills or Model Context Protocol servers. The fake projects borrow the names and workflows of familiar tools, covering Gmail and WhatsApp integrations through to Databricks, Jenkins, and Docker tooling, and their convincing README files walk a user or an agent from routine setup into downloading a malicious ZIP. That triggers a loader chain ending in SmartLoader, which establishes persistence and pulls further payloads such as the StealC infostealer.

Check
Treat AI skills and MCP servers on GitHub as untrusted code: verify the publishing account and project history before installing, and be wary of setup steps that download a ZIP release.
Affected
Developers and AI agents installing skills or MCP servers from GitHub; more than 800 fake repositories impersonate familiar tools, and following their setup instructions leads to SmartLoader and the StealC infostealer.
Fix
Install AI skills and MCP servers only from verified publishers, review repository history and profile age, scan packages before use, and restrict what credentials and systems an installed MCP server can reach.

Fake Go scanning tool hides malware across 222 GitHub repositories

Researchers at Socket uncovered a network of 222 GitHub repositories, tracked as Operation Muck and Load, built to spread Windows malware through a fake Go module posing as a DNS and subdomain scanner. Running the module quietly launches PowerShell that pulls an encrypted payload from attacker infrastructure and deploys infostealers, remote access trojans, and cryptominers. To look trustworthy, the operators used an automated workflow that force-pushes fake commits every minute, making repositories appear actively maintained and generating hundreds of package versions. More than 700 of the module's 1,200-plus versions were malicious. The lures cluster around crypto wallets, Telegram bots, and game cheats, aimed at people likely to run untrusted code.

Check
Review whether developers pulled Go modules or ran tools from unfamiliar GitHub repositories, particularly crypto, wallet, or game-cheat-themed ones, and watch for hidden PowerShell launching from a supposed utility on developer machines.
Affected
Developers and users who clone, build, or run code from these lure repositories; a module posing as a legitimate scanner instead installs infostealers, remote access trojans, or cryptominers on their Windows machines.
Fix
Verify a repository's authenticity beyond how active it looks, since fake commit activity can be manufactured, prefer official module sources, inspect code before running it, and use isolated environments for untrusted tools.

A single GitHub issue can trick AI workflows into leaking private repository data

Researchers at Noma Labs showed that GitHub's new Agentic Workflows, which let an AI agent read issues and act on repositories automatically, can be tricked into leaking private code through nothing more than a public issue. The technique, GitLost, is indirect prompt injection: an attacker opens an ordinary-looking issue in an organization's public repository, buries plain-English instructions in it, and the agent, which often holds a token with read access across the org's repositories, follows them, fetches files from a private repo, and posts the contents in a public comment. No credentials, coding, or write access are needed. GitHub was notified, but researchers frame it as an architectural weakness.

Check
Review any GitHub Agentic Workflows or similar AI automations, and check whether their agents hold organization-wide or cross-repository read tokens while also processing untrusted content like public issues and pull requests.
Affected
Organizations using GitHub Agentic Workflows where the agent has broad cross-repository read access and reacts to public issues; an attacker can steer it into leaking private repository contents with one crafted issue.
Fix
Scope agent tokens to only the repositories a workflow needs, keep agents reading untrusted input away from private data, restrict which events trigger automation, and require human review before agents post publicly.

Clean GitHub repos trick AI coding agents into fetching and running malware

Researchers at Mozilla's 0DIN found that an AI coding agent told to clone and set up a seemingly harmless GitHub repository can be tricked into running malware that stays invisible to security scanners, the agent itself, and human reviewers. The trick is that nothing malicious sits in the repository's files. Instead, a routine-looking setup command runs a script that fetches a value hidden in a DNS TXT record and executes it as a shell command, pulling down and running an attacker's payload like a reverse shell. Because the payload lives outside the repo and arrives over DNS at setup time, code review and static scanning see nothing wrong.

Check
Review how your AI coding agents and developers set up unfamiliar repositories, and check whether setup or build commands can make outbound network or DNS requests that fetch and execute external content.
Affected
Developers and teams that let AI coding agents automatically run setup steps for untrusted repositories; the malicious payload is fetched at setup time over DNS, so scanning the repository alone misses it.
Fix
Run repository setup for untrusted code in sandboxes without credentials, restrict outbound network and DNS during setup, and treat agent setup and build commands as untrusted code execution rather than safe automation.

Miasma worm hits 73 Microsoft GitHub repos, targets AI coding tools

The self-spreading Miasma worm, a variant of the Shai-Hulud malware linked to the group TeamPCP, has reached Microsoft's own code. Using a stolen access token, attackers pushed a malicious commit into the Azure durabletask repository, and GitHub disabled 73 repositories across four Microsoft organizations including Azure and MicrosoftDocs. The twist: the planted code runs automatically when a developer opens the project in an AI coding assistant like Claude Code, Cursor, Gemini CLI, or VS Code, then harvests cloud and developer credentials and uses them to infect more projects. It hides the trigger inside a build file (binding.gyp) that most security tools ignore.

Check
Search your GitHub orgs for commits, public repos, or build files matching Miasma naming patterns, and review AI coding agent configs (binding.gyp, agent rules) for unexpected auto-run payloads.
Affected
Organizations using npm, PyPI, or GitHub alongside AI coding assistants (Claude Code, Cursor, Gemini CLI, VS Code). Stolen maintainer tokens enable backdoored package and repo publishing.
Fix
Rotate GitHub, npm, and cloud credentials exposed to affected projects. Remove malicious commits and configs, enforce 2FA and short-lived tokens, and block install-time scripts in CI.