A dynamic memory curation skill that automatically updates preferences and topics based on historical run feedback.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install cspr-memory-curator
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 cspr-memory-curator 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 Curator?
The Memory Curator skill is designed to automate the evolution of user preferences within a structured retrieval system. By analyzing continuous user interactions, this skill helps keep preferences up to date without manual editing. As a powerful option among Openclaw Skills, it acts as a self-improving personalization layer that bridges the gap between raw feedback data and optimized agent behavior.
Using this skill ensures that agent actions are consistently aligned with refined user taste. It evaluates inputs from various feedback paths and safely scales interest values up or down depending on historical trend lines, preventing sudden, jarring changes to the experience.
Memory Curator Use Cases
- Processing explicit feedback actions such as save, more, less, and hide-source to refine content relevance.
- Adapting to persistent interests dynamically after several output generation cycles.
- Consolidating scattered feedback JSONL entries into an organized, single-source-of-truth configuration.
- Powering tailored content delivery pipelines with automated personalization via Openclaw Skills.
How Memory Curator Works
- Read Feedback Logs: Accesses
~/.cspr/state/feedback.jsonlto parse the log of active feedback entries. - Analyze Outcomes: Scans recently output newspaper content when relevant to detect delivery patterns.
- Load Base Memory: Reads the current reference schema and parameters inside
~/.cspr/memory/profile.yaml. - Apply Updates: Translates the parsed signals into profile modifications (adjusting topics, source weights, and negative filters).
- Stage Profile: Outputs the newly computed preference file securely to the run directory (
RUN/profile.yaml). - Execute Profile Update: Commits the configuration changes to the active memory context using the prefmem CLI.
Memory Curator Setup
To deploy the Memory Curator skill, verify that the required directories and baseline files are present in your workspace.
Ensure that the feedback tracking files are accessible:
# Verify feedback logs
cat ~/.cspr/state/feedback.jsonl
# Inspect current profile config
cat ~/.cspr/memory/profile.yaml
Once files are initialized, run the curation sequence to generate and apply updates via Openclaw Skills workflows:
# Apply the generated configuration staging
prefmem update --in RUN/profile.yaml
Memory Curator Data Schema & Taxonomy
The skill manages preferences across three core file schemas:
| Location | Format | Description |
|---|---|---|
~/.cspr/state/feedback.jsonl |
JSON Lines | Contains individual event logs with actions like save, more, less, and hide-source. |
~/.cspr/memory/profile.yaml |
YAML | The master configuration file tracking source preferences, topic scores, and disliked phrases. |
RUN/profile.yaml |
YAML | Staged intermediary profile file populated during workflow curation before import. |
Memory Curator Advanced Features
- Overfitting Protection: Incorporates conservative updating mechanics designed to prevent dramatic parameter shifts from a single isolated action.
- Action-Based Scoring: Evaluates differing intensities of user intent by mapping save, more, less, and hide-source actions to appropriate weight vectors.
- Direct CLI Integration: Uses isolated staging before executing actual updates to the master memory, allowing for dry-run validation steps in custom Openclaw Skills pipelines.
SKILL.md
Loading
Related Openclaw Skills
Bohrium Dataset Management
An orchestrator for Bohrium datasets supporting resumable multi-gigabyte uploads, multi-version management, and seamless environment mounting.
Juncheng Industrial AI
An AI-powered industrial automation assistant designed to analyze PLC code, design system workflows, diagnose equipment faults, and generate SCADA reports.
Juncheng AI Tutor
A zero-threshold educational skill that guides developers and beginners through the core concepts, principles, and practical application development of Large Language Models.
eqxiu-market-calendar
An automated agent tool for querying Eqxiu marketing holidays and generating ready-to-use digital template marketplace search links.
CritiGraph VM
An open-source, mathematical adversarial reasoning virtual machine that uses Monte Carlo Tree Search (MCTS) to map out opposing arguments and calculate optimal decision paths.
Concept Anatomy
Concept Anatomy uses an intensive 8-dimensional cognitive framework to dissect single concepts from origin to core epiphany.
Featured*
