CVE-2026-105135

InternLM MindSearch Planner Agent graph.py ExecutionAction.run code injection

Published Oct 4, 2026 CVSS 4.0 · 10 Critical EPSS < 1% KEV · Not listed
Description
A vulnerability has been found in InternLM MindSearch 0.1.0. This issue affects the function ExecutionAction.run of the file mindsearch/agent/graph.py of the component Planner Agent. The manipulation of the argument inputs leads to code injection. The attack can be initiated remotely. The exploit has been disclosed to the public and may be used. The vendor was contacted early about this disclosure but did not respond in any way.

Analysis

AI-generated

Technical impact

InternLM MindSearch version 0.1.0 contains a flaw in the ExecutionAction.run method within the Planner Agent component. The function accepts an argument named inputs that can be manipulated to inject arbitrary code. When the input is processed, the injected code is executed, allowing an attacker to run arbitrary commands on the system. The vulnerability is a classic code injection problem exposing the application to total compromise of confidentiality, integrity, and availability through remote execution.

Affected Systems

InternLM MindSearch 0.1.0 is affected. No other product or version information is listed. The vulnerability exists in the Planner Agent portion of the MindSearch system.

Risk and Exploitability

The flaw is rated with a CVSS score of 10, indicating the highest severity. The EPSS score is not disclosed, but the public disclosure and the statement that the attack can be initiated remotely indicate a high likelihood of exploitation. The vulnerability is not yet indexed in CISA's KEV catalog, yet its remote code execution nature warrants immediate attention. Attackers can exploit the flaw by sending crafted inputs to the Planner Agent from a remote location, enabling code execution without additional privileges.

Remediation

No solution or workaround provided in the CVE record.

OpenCVE Recommended Actions

  • Restrict the Planner Agent from processing inputs received from untrusted or external sources; where possible, isolate the agent in a sandboxed environment.
  • Sanitize and validate the "inputs" argument rigorously before it is passed to any code execution path; reject or encode any characters that could form executable code.
  • Run the MindSearch service using the least privilege principle, ensuring that even if code injection occurs, the attacker gains minimal system access.

Generated by OpenCVE AI on October 4, 2026 at 07:20 UTC.

Tracking

Sign in to view the affected projects.

Advisories

No advisories available
History
Type Values Removed Values Added
Metrics ssvc

{'options': {'Automatable': 'yes', 'Exploitation': 'poc', 'Technical Impact': 'total'}, 'version': '2.0.3'}

Type Values Removed Values Added
Description A vulnerability has been found in InternLM MindSearch 0.1.0. This issue affects the function ExecutionAction.run of the file mindsearch/agent/graph.py of the component Planner Agent. The manipulation of the argument inputs leads to code injection. The attack can be initiated remotely. The exploit has been disclosed to the public and may be used. The vendor was contacted early about this disclosure but did not respond in any way.
Title InternLM MindSearch Planner Agent graph.py ExecutionAction.run code injection
First Time appeared Internlm
Internlm mindsearch
Weaknesses CWE-74
CWE-94
CPEs cpe:2.3:a:internlm:mindsearch:*:*:*:*:*:*:*:*
Vendors & Products Internlm
Internlm mindsearch
References
Metrics cvssV2_0

{'score': 10, 'vector': 'AV:N/AC:L/Au:N/C:C/I:C/A:C/E:POC/RL:ND/RC:C'}

cvssV3_0

{'score': 10, 'vector': 'CVSS:3.0/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H/E:P/RL:X/RC:C'}

cvssV3_1

{'score': 10, 'vector': 'CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H/E:P/RL:X/RC:C'}

cvssV4_0

{'score': 10, 'vector': 'CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H/E:P'}

Weaknesses
  • CWE-74

    Improper Neutralization of Special Elements in Output Used by a Downstream Component ('Injection')

  • CWE-94

    Improper Control of Generation of Code ('Code Injection')

Sources

Status
PUBLISHED
Assigner
VulDB
Published
Updated
Oct 6, 2026, 9:54 PM
Reserved
Oct 3, 2026, 1:19 PM
Status
Deferred
Published
Oct 4, 2026, 7:16 AM
Modified
Oct 6, 2026, 10:17 PM