A sophisticated, self-improving agent architecture that shifts AI from reactive task-following to proactive partnership and persistent memory management.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install joko-proactive-agent
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 joko-proactive-agent 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 Joko Proactive Agent?
The Joko Proactive Agent is a battle-tested architecture designed to transform how AI agents interact with users and their environments. Instead of waiting for instructions, this system enables agents to anticipate needs, perform autonomous check-ins, and maintain a persistent 'Soul' and 'User' profile. By integrating this into your Openclaw Skills collection, you provide your agents with the framework necessary to think like owners rather than employees, creating leverage and momentum without constant human intervention.
At its core, the agent utilizes a three-tier memory system and the Write-Ahead Logging (WAL) protocol to ensure that critical details are never lost during context compaction. This makes it an essential tool for developers looking to build resilient, stateful AI workflows that survive long-running sessions and complex task transitions.
Joko Proactive Agent Use Cases
- Maintaining continuity in long-term development projects where context loss is a risk.
- Automating background system maintenance and security audits without user prompts.
- Creating personalized AI assistants that adapt to specific user preferences and 'Soul' principles over time.
- Implementing self-healing technical workflows that attempt multiple resolutions before escalating to a human.
How Joko Proactive Agent Works
- Initialization: The agent detects ONBOARDING.md and populates USER.md and SOUL.md to establish its identity and the user's goals.
- WAL Protocol: The agent scans every input for decisions or corrections, writing them to SESSION-STATE.md before generating a response.
- Context Monitoring: Upon reaching 60% context usage, the Working Buffer Protocol activates to log all exchanges into a 'danger zone' file.
- Compaction Recovery: When context is truncated, the agent automatically reads the working buffer and session state to resume work without asking "where were we?"
- Autonomous Crons: The system utilizes isolated agent turns to perform background tasks like memory distillation and pattern recognition.
Joko Proactive Agent Setup
To deploy this proactive architecture within your Openclaw Skills workspace, execute the following commands:
# Copy the necessary markdown assets to your workspace
cp assets/*.md ./
# Run the initial security audit to vet the environment
./scripts/security-audit.sh
After setup, the agent will recognize the ONBOARDING.md file and guide you through the initial configuration of your preferences and principles.
Joko Proactive Agent Data Schema & Taxonomy
The Joko Proactive Agent organizes data across several Markdown files to ensure high-fidelity memory and architectural integrity:
| File | Role | Update Frequency |
|---|---|---|
SESSION-STATE.md |
Active RAM for current task details and decisions | Every message |
MEMORY.md |
Curated, long-term wisdom distilled from daily logs | Periodically |
USER.md |
Comprehensive context regarding human goals and styles | Dynamic |
SOUL.md |
The agent's core identity, boundaries, and rules | Static |
working-buffer.md |
Raw capture of exchanges in high-context 'danger zones' | >60% Context |
Joko Proactive Agent Advanced Features
- WAL Protocol: Ensures stateful operations by writing critical data before the chat history buffer is flushed.
- Relentless Resourcefulness: A non-negotiable directive to attempt 5-10 alternative approaches before reporting failure.
- ADL/VFM Guardrails: Anti-Drift and Value-First Modification protocols that prevent the agent from adopting complex or useless behaviors.
- Unified Search: A multi-source search strategy that combines semantic search with grep fallbacks to prevent 'I don't know' responses.
- Isolated agentTurn Crons: Allows the agent to perform maintenance tasks autonomously without interrupting the main session flow.
SKILL.md
Loading
Related Openclaw Skills
Autonomous Skill Orchestrator
A deterministic coordination engine that manages autonomous planning, multi-agent delegation, and result verification for complex development tasks.
Joko Moltbook
A specialized integration for AI agents to autonomously engage, post, and monitor activity on the Moltbook social network.
JobHunter Master
An aggressive career-hunting agent that replaces passive applications with a high-conversion, automated sales funnel approach.
Google Flow AI Video Generation Skill
Automate the Google Flow interface to generate professional AI videos from text prompts using the Veo model.
Content Factory
An all-in-one automated YouTube content production suite for generating regular videos and viral Shorts from text or long-form content.
Sales AI Assistant
A comprehensive sales operations engine that transforms AI agents into elite sales partners for lead tracking, outreach, and pipeline management.
METADATA
Featured*
