GitSpawn Exposes Persistent Git Config Execution Risks in Coding Agents
Manifold Security has published GitSpawn, a report uncovering eight vulnerabilities across seven coding agents that allow arbitrary code execution simply by opening a received folder.
The attack leverages Git's core.fsmonitor configuration key stored in .git/config. This setting defines a command that Git automatically runs during index refreshes triggered by operations such as git status or git diff. Coding agents invoke these commands in the background immediately after opening a directory to build context, executing the payload before any user approval or prompt.
The researcher demonstrated the exploit by setting core.fsmonitor to a malicious script and confirming execution with a single git status command. Payloads ran even in the absence of any explicit user action beyond opening the folder.
Partial Patches Leave Broader Risks
Vendors responded with targeted fixes. Claude Code closed the core.fsmonitor vector in version 2.1.196, and similar patches were applied to Codex, Cursor, and Goose. However, these mitigations only prevent the agent from invoking Git with dangerous configuration values. Manual execution of git status or integration with build scripts, editors, and CI systems still triggers the payload.
Additional issues compound the problem. Manifold Security reported a second bypass path in Claude Code that remained open on version 2.1.252. Agents including Qwen Code, Grok Build, and Hermes Agent stayed unpatched as of the report date. A separate vector involving nested bare repositories was assigned CVE-2026-45033 and affected GitHub Copilot CLI until version 1.0.43.
The class of vulnerability is not new. VS Code addressed similar risks in 2021 through workspace trust prompts, yet the arrival of autonomous coding agents effectively reset those protections.
Stroq Tool for Pre-Opening Inspection
To address the gap, the researcher released Stroq, an open-source Apache-2.0 tool that scans repositories for dangerous settings before they are accessed. The command checks .git/config for executable values, .gitattributes, .husky hooks, postinstall scripts, and nested bare repositories.
Environment variable overrides can also neutralize the vectors at runtime by forcing core.fsmonitor to false and disabling automatic bare repository discovery. The tool is intended for use in pre-commit hooks and CI pipelines.
Related articles
Four Characters Bypass SSRF Filter: IPv6 Zone and netip.Prefix.Contains in Go
A vulnerability tracked as CVE-2026-126799 affects Gotenberg versions 8.31.0 through 8.37.0 and allows attackers to bypass SSRF protections by appending an IPv6 zone identifier to addresses such as [::%251]. The flaw stems from how Go's net/netip package handles zone information in methods like Prefix.Contains and IsUnspecified, which ignore or mishandle the zone component. Researchers reproduced the issue on Go 1.26.11 and 1.27.4, showing that addresses with zones like ::%1 or fec0:1234::3%eth0 evade common allow-list checks used in webhooks, link previews, and URL proxies. The root cause lies in documented behavior where netip.Prefix.Contains returns false for any address carrying a zone, while some methods like IsLoopback remain zone-agnostic. Fixes include stripping the zone with WithZone before classification, rejecting zones in URLs outright, and adding a second check inside net.Dialer.Control. The same pattern appears in many Go projects, making the issue broader than a single Gotenberg bug.
Researchers Disclose Full Remote Vehicle Takeover Chain in Major Automaker Connected Platform
Security researchers Shubs and Sam Curry discovered a complete attack chain allowing remote control of any connected vehicle from a major undisclosed automaker across Australia and other large markets. The chain began with decompiled Android APK analysis that revealed hardcoded API endpoints under redacted-telematics.com.au and exposed administrative interfaces. Weak access controls in an OpenIDM instance permitted unauthenticated enumeration of user accounts via email, GUID, and phone number lookups. Path traversal bypasses combined with leaked staging credentials granted administrative access to ForgeRock OpenIDM, enabling password resets and full account takeover. Researchers further bypassed MFA using Resource Owner Password Credentials flows and obtained valid mobile application tokens. With these tokens they could issue arbitrary remote commands including unlock, lock, horn activation, and engine start to any vehicle by VIN or license plate.
Nightmare Eclipse Publishes SignDrop Zero-Day in Windows CLFS Driver
Security researcher Abdelhamid Naseri, known as Nightmare Eclipse, has disclosed a new zero-day vulnerability named SignDrop affecting the Common Log File System (CLFS) driver in Windows. The flaw allows attackers to bypass authentication protections introduced by Microsoft in October 2025 and escalate privileges on affected systems. SignDrop exploits the way CLFS handles .cnpf authentication file creation when a new log file is generated, enabling a privileged process to open the file later and create it with SYSTEM rights. The technique can also be abused against any antivirus or EDR engine to disable self-protection mechanisms by making file creation requests appear to originate from the legitimate security process itself. Nightmare Eclipse confirmed the issue works on Windows 11 26H2, Windows 11 26H2 Experimental, and Windows Server 2025 Experimental, impacting an estimated 80 percent of Windows installations. No workaround exists other than waiting for a Microsoft patch, as disabling CLFS protections would reintroduce memory corruption risks. The researcher previously released multiple other zero-days, including BigDiskBuster and RoguePlanet targeting Microsoft Defender.
Telegram Patches RCE Vulnerabilities in HTML Widgets After Researcher Demonstrates Calculator Exploit
Telegram introduced HTML widgets for bots in early November, allowing JavaScript execution inside messages via the QuickJS engine. Security researcher V12 quickly demonstrated remote code execution by launching the system calculator on Android. Pavel Durov responded that no vulnerability existed because the user had to first start the bot, yet server-side fixes were deployed anyway. Multiple use-after-free and use-after-return bugs were identified in QuickJS and QuickJS-ng, some discovered with AI assistance from Claude Fable 5.1 and Mythos 5.2. Telegram isolated the widget engine into a separate process on Android, added syscall filtering, and limited nested graphical effects to prevent crashes. Desktop clients received the fixes across versions 7.4.2 and 7.4.3, while widgets were temporarily restricted server-side. Users are advised to update to Android 13.2.2 or Desktop 7.4.3 and avoid unknown bots.