An AI agent skill that enables proactive behavior through anticipation, reverse prompting, and self-healing context recovery.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install tmp-proactivity-review
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install tmp-proactivity-review using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Proactivity (Proactive Agent)?
Proactivity transforms your AI agent from a passive command-follower into a high-momentum partner. By maintaining a local state in the proactivity directory, the agent learns to anticipate your next steps, identify hidden blockers, and offer concrete solutions before you even ask. It leverages advanced techniques like reverse prompting and working-buffer recovery to ensure no context is lost during long tasks or system interruptions.
This Openclaw Skills enhancement focuses on relentless resourcefulness and self-correction, allowing agents to diagnose and fix workflow issues independently while respecting strict security boundaries. It is designed for developers and power users who want their agent to think ahead, keep projects moving without constant prompting, and recover context quickly after interruptions.
Proactivity (Proactive Agent) Use Cases
- Accelerating long-running development tasks by maintaining a volatile working buffer for instant context recovery.
- Reducing manual prompting by allowing the agent to suggest next steps, drafts, and potential blockers through reverse prompting.
- Ensuring project momentum by having the agent follow up on stale blockers or missed steps within established boundaries.
- Implementing self-healing workflows where the agent attempts multiple resolution paths before escalating issues to the user.
How Proactivity (Proactive Agent) Works
- Initializes a local state directory at ~/proactivity/ to separate durable rules from active working memory.
- Monitors the current task lifecycle using session-state.md and heartbeat.md to track objectives and next moves.
- Analyzes the workspace and history to identify opportunity signals for proactive intervention or reverse prompting.
- Executes self-healing logic by trying alternative methods or tools before requesting user assistance.
- Updates the patterns.md log with successful proactive moves to refine future behavior and improve over time.
Proactivity (Proactive Agent) Setup
First, ensure the directory structure exists by running the setup guide.
# Create the proactivity directory if it doesn't exist
mkdir -p ~/proactivity/domains
# Install the skill via clawhub
clawhub install proactivity
The skill relies on files like memory.md and heartbeat.md within the ~/proactivity/ folder to manage its state. If the folder is empty, refer to the internal setup.md guide provided within the skill package to initialize your Openclaw Skills environment.
Proactivity (Proactive Agent) Data Schema & Taxonomy
The skill organizes data within ~/proactivity/ using a specific file-based taxonomy:
| File | Description |
|---|---|
| memory.md | Stable activation rules and durable boundary definitions. |
| session-state.md | Tracks current tasks, last decisions, and the immediate next move. |
| heartbeat.md | Records lightweight recurring checks and follow-up items. |
| patterns.md | A library of reusable proactive moves that have successfully worked in the past. |
| working-buffer.md | Volatile breadcrumbs used for context recovery during complex operations. |
Proactivity (Proactive Agent) Advanced Features
- Reverse Prompting: Proactively surfacing ideas and drafts without waiting for user input.
- Self-Healing Logic: Automatic diagnosis and retry mechanisms for broken workflows.
- Context Recovery: Using a working buffer to survive session compactions or interruptions.
- Domain Overrides: Customizable proactive behaviors specific to different project domains.
- Multi-Skill Integration: Seamlessly works with self-improving and heartbeat modules for enhanced Openclaw Skills functionality.
SKILL.md
Loading
Related Openclaw Skills
spec-engine
An automated workflow engine that transforms project ideas into detailed specifications and actionable sub-task lists.
InfoDashboard
A guided SOP skill for deploying InfoDashboard to generate Streamlit visualizations using natural language requirements.
ioc Brand Insights
A professional-grade intelligence skill designed to analyze the market positioning and core product characteristics of the ioc brand.
FIFA Global Football & World Cup Knowledge Skill
A comprehensive data and knowledge skill covering FIFA's organizational structure, World Cup history, and global sports governance.
Booking.com International Hotel Assistant
An expert international travel assistant that enables AI agents to search, compare, and manage Booking.com reservations across 200+ countries.
Booking.com International Hotel Booking Assistant
A comprehensive AI agent skill for global hotel search, price comparison, and reservation management via the Booking.com Affiliate API.
Featured*
