Super Proactive

What This Skill Does

The Super Proactive skill transforms an OpenClaw agent from a reactive chatbot into an autonomous, state-aware collaborator. By implementing a sophisticated architecture that combines a Write-Ahead Logging (WAL) protocol with a three-tier memory structure, this skill ensures your agent maintains context across session flushes. It functions by managing a centralized SESSION-STATE.md file, which serves as a durable workspace buffer, and a task-oriented QUEUE.md for autonomous workflow management. The skill enables the agent to perform scheduled background tasks, such as memory health checks, research updates, and project summarization, ensuring that the agent is always prepared to tackle the next phase of a project without requiring manual reminders.

Installation

To install this skill, use the following terminal command within your OpenClaw environment: clawhub install openclaw/skills/skills/heldinhow/super-proactive. After installation, you must initialize your workspace by creating the necessary directory structure: mkdir -p memory/$(date +%Y-%m-%d) and generating the base SESSION-STATE.md and QUEUE.md files. Ensure you also create or configure a HEARTBEAT.md file to dictate the agent's autonomous task intervals (30-minute checks, 4-hour research windows, and daily cleanups).

Use Cases

This skill is ideal for long-term project management, research-heavy workflows, and technical development. It is particularly effective when you need an agent to track complex dependencies, maintain context over several days of development, or handle repetitive background research. If you are building software and need the agent to track bug fixes, feature requests, and code decisions across multiple sessions, Super Proactive acts as a persistent project lead that logs every critical decision in the session buffer.

Example Prompts

1. "Check the current task queue in QUEUE.md and tell me what the priority is for the next four hours."
2. "Review the entries in SESSION-STATE.md from this morning and summarize the architectural decisions we made for the login module."
3. "Add a new task to the 'Ready' section of the queue: Research the best libraries for real-time data visualization and add findings to MEMORY.md."

Tips & Limitations

For best results, prioritize consistency in how you interact with the files. Always ensure the agent has write permissions for the workspace. Remember that the agent relies heavily on your willingness to let it run background processes; avoid manually overwriting files if the agent is in the middle of a process. Limitations include a dependency on consistent file system access; if the storage layer is unstable, the WAL protocol may fail to sync correctly, leading to potential data loss in the current session state. Regularly audit the daily logs in the memory/ folder to ensure the semantic memory remains relevant and free of noise.