Executive brief
mise is a development tool manager used to handle environments like Node.js and Python. A security flaw allows a malicious project to run unauthorized commands on a user's computer if they simply enter the project directory and list available tasks or use tab-completion. This bypasses the tool's built-in security prompts, potentially allowing an attacker to steal data or install malware when a user interacts with a downloaded repository.
Technical details
A vulnerability exists in mise's task loading logic where task-include directories (e.g., mise-tasks/, .mise/tasks/) are processed without triggering the 'trust_check' mechanism if no primary config file (mise.toml) is present. The application uses the Tera templating engine to render task fields, which includes a registered 'exec()' function. An attacker can place a malicious Tera template, such as {{ exec(command='...') }}, inside a task's description or other fields. These commands are executed automatically when a user runs 'mise tasks', 'mise task ls', or even triggers shell completion (Tab) in the affected directory. The issue is fixed in version 2026.6.4 by ensuring trust checks are applied to all task-include paths.
Affected products
- jdx mise < 2026.6.4
Timeline
- 2026-06-12: advisory: GitHub security advisory published by vendor
- 2026-06-26: disclosed: CVE published to NVD
- 2026-06-26: patched: Fix released in version 2026.6.4