How broadly exploitable is this vulnerability?

Exploitability depends on whether an attacker can influence the directory in which OpenClaw runs. In controlled personal environments this may be limited, but in multi-user systems, shared development environments, or scenarios where projects are opened from untrusted sources, an attacker may be able to supply a crafted directory name and gain influence over the assistant's behavior.

Does an attacker need system-level access to exploit this?

The attacker does not necessarily need full system privileges, but they must be able to create or select a workspace path used by OpenClaw. This could be achieved by sharing a maliciously named project folder, manipulating CI or automation configuration, or convincing a user to clone or unzip content into a directory with a crafted name.

What is the safest general fix pattern for this kind of issue?

The safest approach is to treat all non-literal prompt parameters as untrusted, sanitize or encode them before insertion, and prevent special characters from altering prompt structure. Specifically, strip or normalize control characters, enforce a restricted character set for paths or labels in prompts, and separate data fields from instruction text via structured templates.

Severity:

HIGH RISK

(7.8)

Prompt Injection via Unsanitized Workspace Path (CWE-77) in OpenClaw Personal AI Assistant

CVE-2026-27001

Published:

Feb 19, 2026

Status:

Analyzed

Summary

Affected Context

Exploit Preconditions

Why This Happens

Potential Impact

General Mitigation Guidance

How Teams Detect Issues Like This

Frequently Asked Questions

Source

NDV

Recent Vulnerabilities

CWE-77;CWE-78

HIGH RISK

HIGH

(7.3)

OS Command Injection (CWE-78) in Totolink A7100RU setGameSpeedCfg CGI Handler

Remote Code Execution on the Router

April 6, 2026

CWE-77;CWE-78

HIGH RISK

HIGH

(7.3)

OS Command Injection (CWE-78) in Totolink A7100RU setFirewallType Handler

Remote Code Execution on Router Firmware

April 6, 2026

Vulnerability Type

Category:

Prompt Injection / Command Injection via Control Characters

CWE:

CWE-77

Impact:

Prompt Injection Leading to Unauthorized LLM Actions

Severity:

HIGH RISK

(7.8)

CVSS Version:

3.1

Vulnerability Type

Category:

Prompt Injection / Command Injection via Control Characters

CWE:

CWE-77

Impact:

Prompt Injection Leading to Unauthorized LLM Actions

Severity:

HIGH RISK

(7.8)

CVSS Version:

3.1

Catch This CVE While Code Is Being Written

Detect vulnerable code directly in pull requests, so CVEs are fixed before merge.

Get Started

Ship clean & secure code faster

START 14 DAYS FREE TRIAL

BOOK A DEMO

Made with Love in San Francisco

355 Bryant St. San Francisco, CA 94107, USA

Ask AI for summary of CodeAnt

Made with Love in San Francisco

355 Bryant St. San Francisco, CA 94107, USA

Ask AI for summary of CodeAnt

Made with Love in San Francisco

355 Bryant St. San Francisco, CA 94107, USA

Ask AI for summary of CodeAnt

Prompt Injection via Unsanitized Workspace Path (CWE-77) in OpenClaw Personal AI Assistant

Summary

Affected Context

Exploit Preconditions

Why This Happens

Potential Impact

General Mitigation Guidance

How Teams Detect Issues Like This

Frequently Asked Questions

Source

Recent Vulnerabilities

HIGH RISK

HIGH

(7.3)

OS Command Injection (CWE-78) in Totolink A7100RU setGameSpeedCfg CGI Handler

HIGH RISK

HIGH

(7.3)

OS Command Injection (CWE-78) in Totolink A7100RU setFirewallType Handler

Vulnerability Type

Vulnerability Type

Vulnerability Type

Catch This CVE While Code Is Being Written

Ship clean & secure code faster

Product

Pricing

Company

Legal

Developers

Compare

Product

Pricing

Company

Legal

Developers

Compare

Product

Pricing

Company

Legal

Developers

Compare