CVE-2026-73222

Claude Code Templates: Unauthenticated OS command injection (RCE) in Claude Code Studio server (--studio)

Published Aug 11, 2026 CVSS 3.1 · 8.8 High EPSS < 1% KEV · Not listed
Description
Claude Code Templates is a CLI tool for configuring and monitoring Claude Code. Prior to 1.29.4, the Claude Code Studio server launched by the --studio option in cli-tool/src/sandbox-server.js binds to all interfaces on port 3444, permits cross-origin requests, and requires no authentication. The POST /api/execute endpoint passes the prompt request-body field to executeLocalTask(), and POST /api/install-agent passes the agentName request-body field to a child process. The same unsafe agent field path is reachable from /api/execute through checkAndInstallAgent(). These attacker-controlled values reach child_process.spawn() with shell execution enabled, causing Node.js to construct a shell command in which metacharacters are interpreted. An attacker who can reach the port directly, or who convinces a developer running Studio to visit a malicious website, can execute arbitrary operating-system commands with the developer's privileges and compromise source code, credentials, and local data. This issue is fixed in version 1.29.4.

Analysis

AI-generated

Technical impact

The vulnerability is an unauthenticated operating‑system command injection in the Claude Code Studio server. By sending crafted POST requests to the /api/execute or /api/install-agent endpoints, an attacker can inject shell commands through the prompt or agentName parameters. The injected input is passed to child_process.spawn() with shell enabled, allowing arbitrary commands to run with the privileges of the host system. This flaw results in full compromise of source code, credentials, and any sensitive data stored locally.

Affected Systems

The issue affects the Claude Code Templates CLI tool distributed by davila7. Versions prior to 1.29.4 are vulnerable; the fix was applied in release 1.29.4.

Risk and Exploitability

The CVSS score of 8.8 classifies it as high severity, but the EPSS score of less than 1% indicates a low probability of current exploitation. The vulnerability is not listed in the CISA KEV catalog. Attackers can reach the vulnerable service by contacting port 3444 directly or by tricking a developer running the Studio into visiting a malicious webpage that triggers the vulnerable API calls. Because the server listens on all interfaces and accepts unauthenticated requests, a remote attacker with network access can exploit the flaw immediately, leading to system compromise.

Remediation

No solution or workaround provided in the CVE record.

OpenCVE Recommended Actions

  • Upgrade Claude Code Templates to version 1.29.4 or later to apply the security fix.
  • Restrict network exposure of the Studio server by configuring a firewall or network ACL to only allow trusted hosts to reach port 3444.
  • If upgrading is not yet possible, disable the --studio option in production or configure the server to bind only to localhost, and enforce authentication or disable CORS on the /api/execute and /api/install-agent endpoints.
  • Monitor for unexpected child_process executions and anomalous API traffic to detect potential exploitation attempts.

Generated by OpenCVE AI on August 12, 2026 at 20:33 UTC.

Tracking

Sign in to view the affected projects.

Advisories

Source ID Title
Github GHSA Github GHSA GHSA-79wm-x847-7cvg Claude Code Templates: Unauthenticated OS command injection (RCE) in Claude Code Studio server (--studio)
History
Type Values Removed Values Added
Metrics ssvc

{'options': {'Automatable': 'no', 'Exploitation': 'poc', 'Technical Impact': 'total'}, 'version': '2.0.3'}

Type Values Removed Values Added
First Time appeared Davila7
Davila7 claude-code-templates
Vendors & Products Davila7
Davila7 claude-code-templates
Type Values Removed Values Added
Description Claude Code Templates is a CLI tool for configuring and monitoring Claude Code. Prior to 1.29.4, the Claude Code Studio server launched by the --studio option in cli-tool/src/sandbox-server.js binds to all interfaces on port 3444, permits cross-origin requests, and requires no authentication. The POST /api/execute endpoint passes the prompt request-body field to executeLocalTask(), and POST /api/install-agent passes the agentName request-body field to a child process. The same unsafe agent field path is reachable from /api/execute through checkAndInstallAgent(). These attacker-controlled values reach child_process.spawn() with shell execution enabled, causing Node.js to construct a shell command in which metacharacters are interpreted. An attacker who can reach the port directly, or who convinces a developer running Studio to visit a malicious website, can execute arbitrary operating-system commands with the developer's privileges and compromise source code, credentials, and local data. This issue is fixed in version 1.29.4.
Title Claude Code Templates: Unauthenticated OS command injection (RCE) in Claude Code Studio server (--studio)
Weaknesses CWE-306
CWE-352
CWE-78
References
Metrics cvssV3_1

{'score': 8.8, 'vector': 'CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H'}

Weaknesses
  • CWE-306

    Missing Authentication for Critical Function

  • CWE-352

    Cross-Site Request Forgery (CSRF)

  • CWE-78

    Improper Neutralization of Special Elements used in an OS Command ('OS Command Injection')

Sources

Status
PUBLISHED
Assigner
GitHub_M
Published
Updated
Aug 13, 2026, 1:54 PM
Reserved
Aug 11, 2026, 2:41 PM
Status
Awaiting Analysis
Published
Aug 11, 2026, 7:18 PM
Modified
Oct 24, 2026, 3:10 PM