Clodex – Open-source agentic IDE with governed execution and graph memory (github.com)

Clodex has introduced an open-source agentic development environment, designed to seamlessly integrate persistent AI tasks, coding, terminal operations, and Git functionalities within a single Electron workspace. The core principle of Clodex is the understanding that model outputs are untrusted inputs, establishing authority through explicit policies and user-controlled reviews. Currently in Technical Preview, the platform features isolated runtimes and safeguards for sensitive operations while reducing reliance on prompts for safety—a significant shift toward more secure AI workflows. Clodex's technical architecture is robust, enabling long-running tasks to maintain context, allowing collaboration across diverse tools, and supporting controlled execution within local and cloud environments. Its features include evidence-backed memory, a managed agent runtime, and a rigorous security model that enforces strict access controls and independent authorization for sensitive actions. The integration of features like task history, Git operations, and user-conducted reviews emphasizes collaborative programming while prioritizing security, marking Clodex as a noteworthy advancement in the AI/ML development landscape.