runx

runx is a native governed runtime developed and published by runxhq, designed to execute and manage agent skills, tools, graphs, and packets within a controlled environment. The software serves as foundational infrastructure for agent-based systems, providing the runtime layer in which discrete agent capabilities—referred to as skills—and callable tools can operate under governance, meaning their execution is subject to oversight and control rather than running unmanaged. In addition to skills and tools, runx supports graphs, which represent structured arrangements of connected operations or agent workflows, and packets, which function as the units of data or work moving through that runtime. This combination positions runx within the category of agent infrastructure and runtime software, relevant to developers and organizations building, orchestrating, or deploying autonomous or semi-autonomous agent systems that require a governed execution layer. Typical use cases include running agent skills in a managed context, coordinating tool invocation by agents, executing graph-based agent workflows, and handling the packets that carry information between these components, all within a single native runtime. Because the runtime is described as governed, it is suited to scenarios where control, structure, and oversight of agent behavior are required alongside execution, distinguishing it from unmanaged execution environments. The current version of runx is 0.9.1, indicating a pre-1.0 release that is still progressing toward a stable major version, and the software has been distributed across a total of eight published versions, reflecting an active release history with iterative development by runxhq. Each successive release contributes to the ongoing evolution of the runtime's capabilities for skills, tools, graphs, and packets. Users evaluating runx can consider version 0.9.1 as the latest available build while noting that earlier versions remain part of its release record, which may be relevant for compatibility, testing, or migration purposes when integrating the runtime into existing agent-based projects or infrastructure.

Tags: