A specialized utility to automatically prune, compact, and manage the size of AI agent memory and log files.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install arc-memory-pruner
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 arc-memory-pruner 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 Memory Pruner?
Memory Pruner is an essential tool for developers building long-running agents. It prevents the common problem of unbounded memory growth, where logs and state files eventually consume too much disk space or overwhelm an agent's context window. By implementing a circular buffer for logs and importance-based retention for state, it ensures that your agent remains efficient and cost-effective. Use this skill to keep your Openclaw Skills organized and prevent boot-up delays caused by reading tens of thousands of tokens of stale memory data.
Memory Pruner Use Cases
- Prevent agent state files from exceeding token limits during boot-up.
- Implement log rotation for long-running autonomous processes.
- Clean up stale historical data older than a specific date.
- Monitor memory file sizes and growth rates across different agent instances.
- Safely preview memory cleanup operations using dry-run mode.
How Memory Pruner Works
- The user or system triggers a pruning command via Python to target specific memory files or log directories.
- The utility analyzes the file content based on specified constraints, such as max lines, file count, or date cutoffs.
- For logs, it manages a circular buffer, deleting the oldest files while keeping the most recent history.
- For state files, it compacts the markdown or text content by removing sections that match the pruning criteria.
- It outputs a summary of the operations performed, ensuring your Openclaw Skills maintain a lean data footprint.
Memory Pruner Setup
Ensure you have python3 installed on your system (macOS or Linux).
# Check for python availability
python3 --version
# Run the stats command to see current memory usage across your agents
python3 scripts/memory_pruner.py stats --dir ~/
Memory Pruner Data Schema & Taxonomy
Memory Pruner operates on standard text, markdown, and log files. It does not require a complex database but tracks metadata through file system attributes and line parsing.
| Target Type | Retention Strategy | Metadata Tracked |
|---|---|---|
| State Files (.md) | Max line count or date-based compaction | Line indices, Date patterns |
| Log Directories | Circular buffer (last N files) | Timestamp-based sorting, File count |
| Global Stats | Cumulative size monitoring | Growth rates, Disk usage in bytes |
Memory Pruner Advanced Features
- Circular buffer management for log directories to prevent disk overflow.
- Regex-based compaction to remove specific patterns or outdated sessions from state files.
- Dry run mode for testing pruning logic without modifying production data.
- Size-based enforcement to strictly limit memory files to a specific byte size.
- Integration with other Openclaw Skills to trigger cleanup after specific tasks or daily maintenance.
SKILL.md
Loading
Related Openclaw Skills
Free Worker Dispatch
A smart routing utility that saves money by delegating routine content, research, and coding tasks to free AI models via OpenRouter.
Department Manager
A command-line orchestration layer for delegating tasks and managing specialized AI agent departments like a CEO.
Compliance Checker
Policy-based security assessment and compliance tracking for AI agent skills.
Compliance Audit Trail
Provides an immutable, tamper-evident audit trail to track and verify the actions, decisions, and budget changes of autonomous agents.
Metrics Dashboard
A lightweight operational monitoring tool to track agent health, API usage, and performance metrics through a text-based dashboard.
Security Audit
A unified security assessment tool that chains scanners and verifiers to harden and secure your entire agent skill ecosystem.
Featured*
