Wolli

Wolli is a tool for creating autonomous agents designed to develop around a defined purpose. Each agent begins with no predefined identity or function; users specify its purpose, which the agent records as the first line in a file called SOUL.md. Over time, the agent builds its own identity and capabilities through continued use and interaction. The agents created with Wolli possess several core features. They retain memory across sessions, allowing them to remember information and context from previous interactions. Scheduling is built in, enabling agents to run at specified times or intervals. Agents can also react to external events, such as the opening of an issue, which triggers the agent to wake, assess the situation, and respond accordingly. This event-driven behavior allows agents to triage, flag important matters, and report back as needed. A notable aspect of Wolli agents is their self-extending nature. Agents are capable of authoring and installing their own skills, tools, and integrations as they encounter new tasks or requirements. For example, if an agent needs to interact with GitHub to receive issue notifications, it can create the necessary integration and workflow itself. As agents repeat certain routines, they document these as skills, which further enhances their autonomy and functionality. Wolli is delivered as a command-line interface (CLI) tool, installable via npm. The system emphasizes memory, autonomy, sandboxing, and scheduling as foundational elements for every agent. The evidence does not specify details about pricing, licensing, or the intended audience beyond the implication that users interact with agents through files and conversations. Wolli positions itself within the class of autonomous agent tools focused on adaptability and self-improvement, but does not name any specific standards or competing products.