A specialized skill for distilling AI conversation histories into actionable memory files and concise summaries.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install chat-refiner
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 chat-refiner 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 Chat Refiner?
The Chat Refiner skill is designed to maintain the long-term efficiency of AI agents by pruning redundant or useless exchange history. It focuses on preserving high-value data such as explicit instructions, technical configurations, learned skills, and specific user preferences. By integrating this into Openclaw Skills, developers can prevent context window bloat and ensure that only the most critical decision-making logic and historical context are retained in the agent's memory.
Chat Refiner Use Cases
- Cleaning and summarizing extensive chat transcripts to reduce noise.
- Periodically updating MEMORY.md files from raw session logs.
- Performing memory maintenance during system heartbeats or idle periods.
- Consolidating user requests into structured technical documentation within Openclaw Skills.
How Chat Refiner Works
- Fetch input from sessions_history, existing memory files, or raw chat transcripts.
- Apply refinement principles to filter out casual interactions, heartbeat logs, and repeated tool outputs.
- Extract key disciplines, technical configurations, installed skills, and finalized decisions.
- Generate a consolidated summary and write it to a memory file (e.g., YYYY-MM-DD-summary.md) or update the core MEMORY.md.
- Confirm the number of items refined and provide a link to the cleaned output.
Chat Refiner Setup
Ensure the core memory tools are available in your agent environment. This skill utilizes existing session tracking and file editing tools provided within the Openclaw Skills ecosystem.
# Enable the skill within your agent configuration
openclaw skill enable chat-refiner
Chat Refiner Data Schema & Taxonomy
| Component | Purpose |
|---|---|
| sessions_history | Source transcript data including tool calls and responses. |
| memory/*.md | Date-stamped summary files (e.g., 2024-01-01-summary.md). |
| MEMORY.md | The primary persistent long-term storage for refined instructions and configurations. |
Chat Refiner Advanced Features
- Automated heartbeat maintenance for background memory cleanup.
- Intelligent extraction of disciplines to reinforce agent behavior over time.
- Support for multi-session history distillation into a single unified knowledge base.
- Seamless integration with memory_search for context-aware pruning within Openclaw Skills.
SKILL.md
Loading
Related Openclaw Skills
Arena
A specialized skill for AI agents to compete in live frontend development challenges with automated progress reporting and on-chain rewards.
VAST.ai GPU Rental
A powerful automation tool for searching, renting, and managing high-performance GPU cloud instances via VAST.ai.
Peer Reviewer
An AI-driven academic analysis tool that employs a multi-agent system to simulate rigorous peer reviews and find logical flaws.
OpenClaw Data Importer
An intelligent data ingestion engine that automates the migration of disparate file formats into structured Supabase databases using AI.
DashTask
DashTask is a unified REST API integration for managing tasks, projects, and CRM pipelines through intelligent AI agents.
Document Format Skills
A comprehensive toolkit for diagnosing, repairing, and unifying formatting standards in Word documents.
Featured*
