A self-healing loop for AI agents that diagnoses task failures and converts them into permanent, reusable local skills.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install phoenix-loop
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 phoenix-loop 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 Phoenix Loop?
Phoenix Loop is a sophisticated self-healing mechanism designed to turn agent failures into long-term operational improvements. By implementing a systematic loop of diagnosis and crystallization, this skill ensures that once an agent encounters a blocker or repeated friction, it doesn't just stop; it learns. It identifies failure patterns, extracts root causes, and generates recovery protocols that are stored locally, enhancing the resilience of your Openclaw Skills without compromising data privacy.
The core philosophy of this skill is persistent task completion. Instead of repeating the same errors, the agent utilizes a local feedback loop to rise from failures stronger than before. This privacy-first approach ensures that all learned lessons stay on your machine while providing a robust framework for handling unexpected technical hurdles.
Phoenix Loop Use Cases
- Resolving recurring task blockers by automating solution extraction.
- Building a private library of recovery protocols for specific environment errors.
- Improving agent persistence during complex workflows that require manual configuration updates.
- Automating the post-mortem analysis of failed tasks to prevent future downtime within Openclaw Skills ecosystem.
How Phoenix Loop Works
- Diagnose: The agent scans local memory files to identify failed patterns and recent solution attempts using PowerShell diagnostics.
- Extract: It identifies reusable logic from the failure, defining triggers, root causes, and specific fix steps.
- Crystallize: A new recovery skill is written as a Markdown file in the local library to be used by future Openclaw Skills.
- Verify: Upon encountering a similar issue, the agent searches the local library, executes recovery steps, and logs results to update the skill logic.
Phoenix Loop Setup
To get started with this self-healing protocol, ensure your agent has access to the local memory directories and the ability to execute file system commands.
# Ensure the local skills directory exists
mkdir -p skills/local/
# Initialize the memory files if they do not exist
touch memory/blocked-items.md
touch memory/tasks.md
You should also configure your HEARTBEAT.md to include a daily self-check, allowing the Phoenix Loop to trigger its diagnosis phase every 24 hours.
Phoenix Loop Data Schema & Taxonomy
The skill manages data across primary local directories to ensure all Openclaw Skills benefit from learned lessons while maintaining a clear metadata taxonomy.
| Path | Purpose | Format |
|---|---|---|
memory/blocked-items.md |
Active blockers and attempted solutions | Markdown |
memory/tasks.md |
Historical task data for pattern recognition | Markdown |
skills/local/*.md |
Generated recovery protocols and triggers | Markdown |
memory/{date}.md |
Daily execution logs and verification checks | Markdown |
Phoenix Loop Advanced Features
- Automated Privacy Scrubbing: Automatically detects and removes sensitive data like API keys, secrets, and personal information before writing to disk.
- Heartbeat Integration: Supports scheduled 24-hour self-checks to periodically optimize the local skill library without manual intervention.
- Fallback Execution: Every crystallized skill includes mandatory fallback actions to ensure the agent remains responsive even if recovery fails.
- Dynamic Skill Management: Enables power users to disable or roll back generated skills via simple PowerShell command integrations.
SKILL.md
Loading
Related Openclaw Skills
Model Route Guard
A specialized diagnostic and repair tool designed to eliminate model routing conflicts and ensure consistent provider endpoint configurations.
Memory Self-Heal
A resilient self-healing skill that empowers AI agents to diagnose, retry, and learn from execution failures autonomously.
Instruction Anchor Guard
A persistence layer for AI agents that ensures critical user constraints and high-priority rules are never lost during context compaction or long sessions.
Gateway Token Doctor
A diagnostic and repair utility to eliminate 401 Unauthorized errors by synchronizing tokens across all service layers.
PowerShell Reliable Execution
A professional framework for executing robust, error-tolerant PowerShell commands on Windows systems.
PowerShell Safe Chain
A standardized approach for chaining Windows PowerShell commands reliably using try/catch logic and error handling instead of brittle bash-style operators.
Featured*
