InternLM MindSearch: Code injection in Planner Agent
Summary
MindSearch 0.1.0 contains a code injection flaw in the Planner Agent (ExecutionAction.run in mindsearch/agent/graph.py). Manipulated input can lead to code being executed on the server, and the attack can be started remotely. According to the linked exploit description, the POST /solve endpoint can be reached without authentication, so an attacker with network access to the service could run operating system commands with the privileges of the service process. The vendor did not respond to early disclosure, and the exploit is reported as publicly disclosed.
Affected
- InternLM MindSearch 0.1.0 (per VulDB and the sources reviewed)
- Other versions: Unknown.
Fixed in
- Unknown. No vendor advisory or fixed version found.
Workaround
- Remove public routing to MindSearch and restrict network access to trusted internal addresses using firewall rules or a reverse proxy.
- Place the service behind VPN or authenticated access.
- Run the service with minimal privileges in a container without access to secrets or sensitive file systems.
Exploitation
- Public exploit: VulDB states the exploit has been disclosed to the public and may be used. One aggregator lists no public PoC, so this conflicts and is not fully verified.
- Known exploited (KEV): No evidence found.
- CISA SSVC: Exploitation poc, automatable yes, technical impact total.
Our recommendation
- Check whether MindSearch is in use and whether it is reachable from the internet or less-trusted networks.
- Isolate or shut down affected instances until a fixed version is confirmed by the vendor.
- Restrict access with VPN or authentication and monitor logs for unexpected process execution.
- Replace the product or update to a verified fixed release once published.
Research sources
- VulDB CVE-2026-105135 vuldb.com
- Linked exploit description (GitHub Gist) gist.github.com
- HOL Guard CVE entry hol.org
- Strix CVE entry www.strix.ai
- CVE Brief bulletin 2026-10-04 cvebrief.com
- TheHackerWire CVE-2026-105135 www.thehackerwire.com
Researched and written up in English and German by an AI (Claude by Anthropic) based on NVD, CISA, EPSS and the sources listed. The vendor information is authoritative. Researched on 31 Oct 2026, 03:32.
Original description (NVD)
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.
Affected? We can help.
We check whether your systems are affected, install updates and close the gap – continuously with the System Health Checker if you wish.
Data: NVD (NIST), CISA KEV, FIRST EPSS – updated continuously.