FoundDream miniclawd exec.ts ExecTool.execute os command injection
Description
A security vulnerability has been detected in FoundDream miniclawd up to 2d65665046e2222eeea76cafc8570ed546a8c125. Affected by this issue is the function ExecTool.execute of the file /src/tools/exec.ts. Such manipulation leads to os command injection. The attack can be launched remotely. The exploit has been disclosed publicly and may be used. This product does not use versioning. This is why information about affected and unaffected releases are unavailable. The project was informed of the problem early through an issue report but has not responded yet.
AI Insight
LLM-synthesized narrative grounded in this CVE's description and references.
CVE-2026-9452 describes an unpatched OS command injection in FoundDream miniclawd's ExecTool.execute, allowing remote attackers to execute arbitrary shell commands.
Vulnerability
CVE-2026-9452 is an OS command injection vulnerability in FoundDream miniclawd, a lightweight TypeScript-based AI assistant. The flaw resides in the ExecTool.execute function in /src/tools/exec.ts (lines 47–51). The vulnerable code uses Node.js spawn() with shell: true, which passes the entire params.command string directly to /bin/sh for parsing with no input validation. Affected versions include all commits up to 2d65665046e2222eeea76cafc8570ed546a8c125; the project does not use versioning [1][2].
Exploitation
An attacker can launch the attack remotely by sending a crafted message to the AI agent (e.g., via Telegram, Feishu, or CLI). The AI constructs an exec_tool call with a command containing shell metacharacters. For example, an attacker sends: "Run ls; cat /etc/passwd to check system". The AI then executes exec_tool({ command: "ls; cat /etc/passwd" }). Due to shell: true, the shell parses ; as a command separator and executes both commands: /bin/sh -c "ls; cat /etc/passwd". All shell operators (;, |, &, &&, ||, $(), backticks, newlines) work as intended [2].
Impact
Successful exploitation allows an attacker to execute arbitrary OS commands on the server running miniclawd. This can lead to full compromise of the system, including reading sensitive files (e.g., /etc/passwd), modifying data, or leveraging the host for further attacks. The impact is high in terms of confidentiality, integrity, and availability [2].
Mitigation
As of the publication date (2026-05-25), no official fix has been released by the project maintainers. The project was informed via an issue report but has not responded [2]. Users should consider disabling the ExecTool or modifying the code to remove shell: true and validate/sanitize all command input. If the software is used in an untrusted environment, isolate it from sensitive systems until a patch is available. The project does not use versioning, so manual code review and patching are required [1][2].
AI Insight generated on May 25, 2026. Synthesized from this CVE's description and the cited reference URLs; citations are validated against the source bundle.
Affected products
1Patches
0No patches discovered yet.
Vulnerability mechanics
AI mechanics synthesis has not run for this CVE yet.
References
4- github.com/FoundDream/miniclawd/issues/1mitreexploitissue-tracking
- vuldb.com/submit/813767mitrethird-party-advisory
- vuldb.com/vuln/365433mitrevdb-entrytechnical-description
- vuldb.com/vuln/365433/ctimitresignaturepermissions-required
News mentions
0No linked articles in our index yet.