GHSA-PJ5X-38RW-6FPH

Vulnerability from github – Published: 2026-03-03 21:50 – Updated: 2026-03-20 21:35
VLAI?
Summary
OpenClaw has a Command Injection via unescaped environment assignments in Windows Scheduled Task script generation
Details

Summary

A command injection vulnerability existed in Windows Scheduled Task script generation for OpenClaw. Environment values were written into gateway.cmd using unquoted set KEY=VALUE, which allowed Windows shell metacharacters in config-provided environment variables to break out of assignment context.

Affected Packages / Versions

  • Package: openclaw (npm)
  • Affected versions: <= 2026.2.17
  • Patched version: >= 2026.2.19
  • Latest published vulnerable version at review time (2026-02-19): 2026.2.17

Practical Risk Context

For a single-user, localhost-only setup on a personally controlled machine, practical risk is typically low.

This issue becomes materially relevant when configuration or environment values are sourced from less-trusted inputs, for example: - shared/team config templates, - copied config snippets, - setup scripts, automation, or repos that write config, - any workflow where another party can influence env values before gateway install/reinstall.

In those scenarios, it provides a reliable config-to-command-execution path when the scheduled task script is generated and run.

Details

On Windows, gateway service installation writes a helper batch script and then registers it via Scheduled Task (schtasks). Before the fix, env lines were rendered as set KEY=VALUE in src/daemon/schtasks.ts, so values containing metacharacters (for example &, |, ^, %, !) could alter command behavior in cmd.exe.

The fix now renders quoted assignments (set "KEY=VALUE") with explicit escaping for cmd metacharacters, updates parser compatibility for quoted assignments, and adds regression tests for metacharacter handling and round-trip parsing.

Fix Commit(s)

  • dafe52e8cf1a041d898cfb304a485fa05e5f58fb

OpenClaw thanks @tdjackey for reporting.

Show details on source website

{
  "affected": [
    {
      "database_specific": {
        "last_known_affected_version_range": "\u003c= 2026.2.17"
      },
      "package": {
        "ecosystem": "npm",
        "name": "openclaw"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "2026.2.19"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-22176"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-78"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-03-03T21:50:05Z",
    "nvd_published_at": null,
    "severity": "HIGH"
  },
  "details": "### Summary\nA command injection vulnerability existed in Windows Scheduled Task script generation for OpenClaw. Environment values were written into `gateway.cmd` using unquoted `set KEY=VALUE`, which allowed Windows shell metacharacters in config-provided environment variables to break out of assignment context.\n\n### Affected Packages / Versions\n- Package: `openclaw` (npm)\n- Affected versions: `\u003c= 2026.2.17`\n- Patched version: `\u003e= 2026.2.19`\n- Latest published vulnerable version at review time (2026-02-19): `2026.2.17`\n\n### Practical Risk Context\nFor a single-user, localhost-only setup on a personally controlled machine, practical risk is typically low.\n\nThis issue becomes materially relevant when configuration or environment values are sourced from less-trusted inputs, for example:\n- shared/team config templates,\n- copied config snippets,\n- setup scripts, automation, or repos that write config,\n- any workflow where another party can influence env values before `gateway install`/reinstall.\n\nIn those scenarios, it provides a reliable config-to-command-execution path when the scheduled task script is generated and run.\n\n### Details\nOn Windows, gateway service installation writes a helper batch script and then registers it via Scheduled Task (`schtasks`).\nBefore the fix, env lines were rendered as `set KEY=VALUE` in `src/daemon/schtasks.ts`, so values containing metacharacters (for example `\u0026`, `|`, `^`, `%`, `!`) could alter command behavior in `cmd.exe`.\n\nThe fix now renders quoted assignments (`set \"KEY=VALUE\"`) with explicit escaping for cmd metacharacters, updates parser compatibility for quoted assignments, and adds regression tests for metacharacter handling and round-trip parsing.\n\n### Fix Commit(s)\n- `dafe52e8cf1a041d898cfb304a485fa05e5f58fb`\n\nOpenClaw thanks @tdjackey for reporting.",
  "id": "GHSA-pj5x-38rw-6fph",
  "modified": "2026-03-20T21:35:48Z",
  "published": "2026-03-03T21:50:05Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/openclaw/openclaw/security/advisories/GHSA-pj5x-38rw-6fph"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-22176"
    },
    {
      "type": "WEB",
      "url": "https://github.com/openclaw/openclaw/commit/dafe52e8cf1a041d898cfb304a485fa05e5f58fb"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/openclaw/openclaw"
    },
    {
      "type": "WEB",
      "url": "https://www.vulncheck.com/advisories/openclaw-command-injection-via-unescaped-environment-variables-in-windows-scheduled-task"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:4.0/AV:L/AC:L/AT:P/PR:N/UI:A/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N",
      "type": "CVSS_V4"
    }
  ],
  "summary": "OpenClaw has a Command Injection via unescaped environment assignments in Windows Scheduled Task script generation"
}


Log in or create an account to share your comment.




Tags
Taxonomy of the tags.


Loading…

Loading…

Loading…

Sightings

Author Source Type Date

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or observed by the user.
  • Confirmed: The vulnerability has been validated from an analyst's perspective.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
  • Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
  • Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
  • Not confirmed: The user expressed doubt about the validity of the vulnerability.
  • Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.


Loading…

Detection rules are retrieved from Rulezet.

Loading…

Loading…