CVE-2026-7058

7.3

666ghj · MiroFish

MiroFish versions up to 0.1.2 contain a cross-tenant command injection vulnerability in the IPC mechanism, allowing unauthenticated attackers to remotely control or shut down simulation environments.

Executive summary

An unauthenticated remote command injection vulnerability in MiroFish allows attackers to manipulate or terminate arbitrary simulation environments across user tenants.

Vulnerability

The application fails to validate the simulation_id parameter when deriving paths for Inter-Process Communication (IPC) command files in the SimulationIPCClient.send_command function. This allows an unauthenticated attacker to inject arbitrary command files into any directory, which are then executed by simulation subprocesses without origin verification.

Business impact

Successful exploitation allows an attacker to gain unauthorized control over simulation environments, leading to the forced termination of victim processes, data manipulation via prompt injection, and potential cross-tenant disruption. Given the CVSS score of 7.3 and the nature of the vulnerability, this poses a significant risk to the integrity and availability of simulation services, potentially resulting in operational downtime and compromised research outcomes.

Remediation

Immediate Action: As no official patch is currently available, administrators should restrict network access to the backend API to trusted sources only and disable the simulation runner service if possible.

Proactive Monitoring: Monitor access logs for suspicious POST requests to /api/simulation/ endpoints and review filesystem logs for unauthorized command file creation within simulation directories.

Compensating Controls: Deploy a Web Application Firewall (WAF) with custom rules to inspect and block requests containing path traversal patterns or unexpected simulation_id values targeting the IPC service.

Exploitation status

Public Exploit Available: Yes — a public proof-of-concept exists as detailed in the GitHub issue report 488.

Analyst recommendation

This vulnerability represents a high-risk security flaw that enables unauthorized remote control of simulation infrastructure. Organizations utilizing MiroFish 0.1.2 or earlier must prioritize the implementation of the suggested compensating controls and monitor for signs of unauthorized API interaction. Please maintain close watch on the official repository for the release of a security patch and apply it immediately upon availability.

Sources

Originally found and disclosed by York Shen (VulDB User), with VulDB CNA Team (coordinator), per the CVE Program record.