A structured personal memory system that allows AI agents to persist, consolidate, and recall episodic and semantic memories across multiple sessions.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install autobiographical-memory
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 autobiographical-memory 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 Autobiographical Memory?
Autobiographical Memory provides a dual-layered architecture for AI agents to maintain continuity. By distinguishing between episodic memory (raw daily events and conversations) and semantic memory (curated knowledge and user preferences), the skill ensures that agents remain contextually aware over long durations. This Openclaw Skills integration transforms a standard LLM into a persistent collaborator that learns from every interaction.
The system utilizes a robust lifecycle of capturing data in daily logs, consolidating that data into a long-term MEMORY.md file, and performing intelligent recall. This prevents the loss of critical context such as past decisions, rationale for specific workflows, and evolving user preferences, making the agent significantly more effective in complex, multi-day projects.
Autobiographical Memory Use Cases
- Recording significant architectural decisions and the rationale behind them to avoid repeating past mistakes.
- Tracking evolving user preferences and identity changes across long-term interactions.
- Recalling specific context from past conversations or project milestones before starting new tasks.
- Consolidating fragmented daily logs into a high-level knowledge base for streamlined agent performance.
- Managing technical environment quirks and tool configurations discovered during development sessions.
How Autobiographical Memory Works
- Capture: The agent records significant events, decisions, and observations into daily episodic files located in the memory/ directory.
- Search: Before responding to queries, the agent performs a memory_search to retrieve relevant past context from both daily notes and the core MEMORY.md file.
- Consolidate: Periodically, the agent reviews recent daily logs to extract high-value information and promotes it to the semantic MEMORY.md file.
- Review: Stale or obsolete information is pruned from the long-term memory to maintain a high signal-to-noise ratio.
- Recall: The agent synthesizes retrieved memories with current inputs to provide deeply informed and personalized responses.
Autobiographical Memory Setup
To implement this within your Openclaw Skills environment, ensure your agent has access to the following directory structure and scripts:
# Create the memory directory structure
mkdir -p memory/
touch MEMORY.md
# Run the consolidation script to process recent logs
python3 scripts/consolidate.py
# Check memory health and statistics
python3 scripts/stats.py
Autobiographical Memory Data Schema & Taxonomy
The skill organizes data into two primary Markdown-based layers to balance detail with retrieval speed:
| Layer | File Path | Content Type |
|---|---|---|
| Episodic | memory/YYYY-MM-DD.md |
Raw daily events, specific conversation summaries, and raw decisions. |
| Semantic | MEMORY.md |
Curated user preferences, active project status, and long-term lessons learned. |
Standard Metadata Sections:
- Events: Descriptions of what happened.
- Decisions: Actions taken plus the underlying rationale.
- Observations: Patterns or insights noticed by the agent.
- Technical Environment: Discovered tool quirks and configurations.
Autobiographical Memory Advanced Features
- Automated Consolidation: Use the provided Python scripts to automatically scan daily notes and suggest updates to the long-term semantic memory.
- Multi-Agent Context Sharing: Shared memory files allow different agents within the Openclaw Skills ecosystem to maintain a unified understanding of the user.
- Intelligent Forgetting Curve: Implements a systematic review process to archive obsolete data, preventing context window bloat.
- Cross-Reference Support: Integrates with identity files like SOUL.md and USER.md for holistic persona management.
- Heuristic Search Patterns: Supports multiple query phrasings and deep-dive retrieval for specific date ranges.
SKILL.md
Loading
Related Openclaw Skills
Course Syllabus Decoder
An intelligent academic workflow that converts raw course syllabus data into structured semester plans and study systems.
Freelance Client Boundary Kit
A comprehensive automation toolkit for freelancers to manage project scope, handle revisions, and maintain professional boundaries with clients.
Time-Travel TV
A virtual CRT television skill that allows users to 'live stream' any moment in history or the future with deep sensory immersion.
Roast Chef
An AI stand-up comedian that delivers brutal yet creative roasts on code, writing, photos, and life choices.
Designer: Aesthetic System Framework
A sophisticated design system that unifies major art and design movements into a functional framework for creating visual identities and AI prompts.
Backlink Strategy
A comprehensive framework for ecommerce stores to build domain authority through high-value outreach, partnership identification, and the creation of linkable content assets.
Featured*
