Proactivity

What This Skill Does

The Proactivity skill transforms your OpenClaw agent from a reactive tool into an autonomous partner. It enables the agent to anticipate user needs, surface hidden opportunities, and execute tasks independently within predefined boundaries. By utilizing a tiered autonomy framework (DO, SUGGEST, ASK, NEVER), the agent learns your preferences over time, ensuring it acts when you are busy but defers to you for high-impact decisions. The system is architecture-centric, using a structured directory at ~/proactivity/ to store memory, domain-specific rules, and historical patterns, preventing the agent from repeating the same questions and ensuring a consistent, learning-based workflow.

Installation

To integrate this functionality into your environment, run the following command in your terminal: clawhub install openclaw/skills/skills/ivangdavila/proactivity Once installed, ensure the ~/proactivity/ directory is initialized using the provided memory-template.md to establish your baseline autonomy levels.

Use Cases

• Automated Research: When you start a project, the agent proactively gathers background context and summarizes documentation without waiting for explicit prompts.
• Pipeline Maintenance: For developers, the agent can monitor system health, detect failing tests, and suggest specific patches before you even notice the error.
• Workflow Optimization: If you frequently perform repetitive manual data entry, the agent observes the pattern and asks permission to automate the entire loop.
• Contextual Digests: Instead of notifying you for every minor event, the agent batches non-urgent tasks into morning and evening summaries, reducing noise.

Example Prompts

1. "I am starting a new research project on decentralized finance; please use the proactivity skill to start monitoring and summarizing relevant news."
2. "Review the current boundaries in my ~/proactivity/ folder and suggest which tasks I should move from 'ASK' to 'DO' level."
3. "I feel like I am getting too many alerts lately; please review my log.md and throttle non-essential proactive actions by 50%."

Tips & Limitations

To get the most out of Proactivity, be diligent with your initial boundary setting. The "Relentless Execution" rule means your agent will attempt to solve problems using up to 10 different approaches before ever bothering you; ensure your tools are correctly configured. Remember that the system relies heavily on the log.md file for refinement, so conduct weekly reviews to audit what the agent has been doing. The primary limitation is that it requires a clean, structured directory; if your memory.md becomes cluttered, the agent’s decision-making process will degrade in quality.