Microsoft fixes critical VS Code Workspace Trust bypass CVE-2026-81376 (CVSS 9.6); researcher awarded $4,000
Microsoft awarded researcher Sam Chen (Samchen) US$4,000 for reporting CVE-2026-81376, a critical (CVSS 3.1 score 9.6) Workspace Trust security feature bypass in Visual Studio Code versions before 1.136.2. A specially prepared workspace could connect to an attacker-controlled service and bypass Restricted Mode, potentially exposing local data or running code with the logged-in user's permissions. Microsoft has patched the flaw. Its advisory was published on September 8 and the researcher disclosed publicly on September 10. The article reports no in-the-wild exploitation and no malware, and the full technical write-up and proof of concept are not yet public.
Technical Details
- Microsoft describes the weakness as "incomplete comparison with missing factors"; NVD lists it as a network-triggerable security feature bypass. The CVSS vector requires user interaction and rates confidentiality, integrity and availability impact as high.
- Chen says the bug could be used to reach remote code execution. The exact exploit steps have not been disclosed.
- Chen reported the issue to MSRC on May 23, 2026. MSRC confirmed it on June 3, and on June 9 it was found eligible for the Open Source Bounty Program, classified as Important, High Quality and Improper Input Validation.
- NVD lists VS Code from 1.0.0 up to (excluding) 1.136.2 as affected.
- Developer workstations often hold source code, signing keys, cloud credentials and access tokens, which raises the impact of a successful exploit.
Remediation
- Update VS Code to 1.136.2 or later and confirm the editor has restarted on the new build.
- Check managed and developer-owned devices, including machines with more than one VS Code installation.
- Keep Workspace Trust enabled and use Restricted Mode for unfamiliar projects. Be cautious with projects received via unsolicited links, code tests or unknown repositories.
- Review unusual activity on developer endpoints, and do not rely on speculative exploit details or indicators.
Linha do tempo
-
Microsoft fixes critical VS Code Workspace Trust bypass CVE-2026-81376 (CVSS 9.6); researcher awarded $4,000
ler a análise desta fonte ocultar
Microsoft awarded researcher Sam Chen (Samchen) US$4,000 for reporting CVE-2026-81376, a critical (CVSS 3.1 score 9.6) Workspace Trust security feature bypass in Visual Studio Code versions before 1.136.2. A specially prepared workspace could connect to an attacker-controlled service and bypass Restricted Mode, potentially exposing local data or running code with the logged-in user's permissions. Microsoft has patched the flaw. Its advisory was published on September 8 and the researcher disclosed publicly on September 10. The article reports no in-the-wild exploitation and no malware, and the full technical write-up and proof of concept are not yet public.
Technical Details
- Microsoft describes the weakness as "incomplete comparison with missing factors"; NVD lists it as a network-triggerable security feature bypass. The CVSS vector requires user interaction and rates confidentiality, integrity and availability impact as high.
- Chen says the bug could be used to reach remote code execution. The exact exploit steps have not been disclosed.
- Chen reported the issue to MSRC on May 23, 2026. MSRC confirmed it on June 3, and on June 9 it was found eligible for the Open Source Bounty Program, classified as Important, High Quality and Improper Input Validation.
- NVD lists VS Code from 1.0.0 up to (excluding) 1.136.2 as affected.
- Developer workstations often hold source code, signing keys, cloud credentials and access tokens, which raises the impact of a successful exploit.
Remediation
- Update VS Code to 1.136.2 or later and confirm the editor has restarted on the new build.
- Check managed and developer-owned devices, including machines with more than one VS Code installation.
- Keep Workspace Trust enabled and use Restricted Mode for unfamiliar projects. Be cautious with projects received via unsolicited links, code tests or unknown repositories.
- Review unusual activity on developer endpoints, and do not rely on speculative exploit details or indicators.