Arbitrary Code Execution in Language Servers for AWS Workspace MCP Configuration Handling
CVE-2026-12957 is a high-severity arbitrary code execution vulnerability in Language Servers for AWS before version 1.65.0 on all supported platforms, affecting the runtime used by Amazon Q Developer integrations. The issue is caused by improper trust boundary enforcement when processing workspace/project configuration files, specifically reported in connection with Model Context Protocol (MCP) server configuration such as .amazonq/mcp.json. When a user opens a maliciously crafted workspace and trusts it when prompted, the language server may automatically load and execute commands defined in project configuration files. Reported analyses indicate that attacker-defined MCP server processes can be launched locally and inherit the developer’s environment, including credentials and tokens present in the session.
Are you exposed to this one?
Mallory correlates every CVE against your assets, your vendors, and active adversary campaigns. Know which vulnerabilities matter for you, not just which ones are loud.
Impact, mitigation & remediation
What it means. What to do now. Patch path, mitigations, and the assume-compromise checklist.
Impact
What an attacker gets, and what they’ve been doing with it.
Mitigation
If you can’t patch tonight, do this now.
Remediation
Patch, then assume compromise.
Exploits
No public exploits tracked yet. Mallory keeps watching.
No public exploit code observed for this vulnerability.
Affected products & vendors
Products and vendors Mallory has correlated with this vulnerability. Open in Mallory to drill down to specific CPE configurations and version ranges.
Vendor-confirmed product mapping. Mallory continuously reconciles this list against your asset inventory.
Recent activity
13 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.
A high-severity vulnerability in Amazon Q Developer plugins/language server where workspace-embedded configuration files could be auto-executed without user permission, enabling attacker-controlled command execution and theft of cloud credentials and API keys from a developer environment.
A high-severity flaw in Amazon Q Developer / Language Servers for AWS that allowed a malicious repository to define an MCP server via .amazonq/mcp.json, leading to command execution and theft of developer cloud credentials after the workspace was trusted.
A high-severity arbitrary command execution vulnerability in Amazon Q for Visual Studio Code caused by automatically loading and executing MCP server configurations from a repository's .amazonq/mcp.json file without user consent or workspace trust checks, potentially exposing cloud credentials and other secrets.
An arbitrary code execution vulnerability in Language Servers for AWS caused by improper trust boundary enforcement. A local user who opens and trusts a maliciously crafted workspace may trigger automatic execution of commands from project configuration files.
The version that knows your environment.
Query your assets running an affected version, and investigate the blast radius.
Every observed campaign linking this CVE to a named adversary.
Malware families riding this exploit, with evidence and IOCs.
YARA, Sigma, Snort, and vendor rules, auto-deployed to your SIEM.
Cross-references every affected SKU, including bundled OEM variants.
Community discussion across Reddit, Mastodon, and other social sources.