A token-efficient memory architecture that utilizes indexed progressive disclosure to minimize context rot and maximize agent relevance.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install progressive-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 progressive-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 Progressive Memory?
Progressive Memory is a strategic framework designed to combat context rot and token waste in AI agent workflows. By applying the principle of progressive disclosure, this system allows agents to first scan a lightweight index of historical data rather than loading thousands of irrelevant tokens into the prompt context. This approach ensures that the agent's attention budget is preserved for the specific task at hand, making Openclaw Skills significantly more cost-effective and accurate during long-term projects and complex development sessions.
Progressive Memory Use Cases
- Reducing token consumption in complex, long-running agentic workflows.
- Managing technical debt and project gotchas without flooding the context window.
- Streamlining handoffs between different AI agents within the same environment.
- Improving decision-making efficiency for Openclaw Skills by prioritizing high-impact rules and architectural decisions.
How Progressive Memory Works
- At the start of a session, the agent scans a compact index table of memory entries, usually consuming fewer than 200 tokens.
- The agent uses standardized icons to visually categorize entries into rules, pitfalls, decisions, or changes.
- Based on the current task, the agent evaluates the token cost listed in the index to determine the ROI of fetching the full detail.
- The agent fetches only the specific Markdown blocks required by referencing unique IDs like #1 or G3 to maintain high context relevance.
Progressive Memory Setup
To integrate this memory system into your Openclaw Skills setup, follow these steps:
- Update your agent's system instructions or AGENTS.md to recognize the Progressive Memory index format.
- Initialize your long-term memory file:
touch MEMORY.md
- Set up a directory for daily logs:
mkdir -p memory/
- Restructure your existing documentation into the indexed table format to allow for on-demand fetching by the agent.
Progressive Memory Data Schema & Taxonomy
The system utilizes a dual-file hierarchy to organize information efficiently:
| File Type | Purpose | Key Metadata |
|---|---|---|
| MEMORY.md | Long-term rules and architectural decisions | ID, Category, Token Count |
| memory/YYYY-MM-DD.md | Daily session logs and immediate task history | ID, Type Icon, Summary, Tokens |
Each entry follows a strict taxonomy using icons like 🚨 for rules, 🔴 for pitfalls (gotchas), and 🟣 for architectural decisions, allowing Openclaw Skills to scan context at a glance and fetch relevant data blocks.
Progressive Memory Advanced Features
- Token Estimation: Every entry includes an estimated token count, enabling proactive context management and cost awareness.
- Visual Pattern Recognition: Standardized emoji icons help agents prioritize critical information types like rules and fixes over general changes.
- On-Demand Retrieval: Agents can peek at the index and selectively load high-value blocks, significantly reducing daily token overhead.
- Markdown-Native Design: Fully compatible with existing Markdown-based Openclaw Skills and file-system-driven agent protocols.
SKILL.md
Loading
Related Openclaw Skills
Prisma Generator
A command-line utility that transforms natural language descriptions into complete, syntactically correct Prisma database schemas.
Prisma Gen
Prisma Gen transforms natural language descriptions into fully formatted, production-ready Prisma database schemas in seconds.
Prefetch Advisor
Prefetch Advisor leverages AI to analyze application routing and user flows, recommending specific data and routes to prefetch for instantaneous navigation.
Prefetch Suggester
An AI-powered CLI tool that analyzes application routes to suggest and implement intelligent prefetching for instant page transitions.
Query Optimizer
An AI-driven CLI tool designed to identify and fix performance bottlenecks in SQL and Prisma database queries.
Queue Generator (queue-gen)
A command-line utility that instantly generates production-ready BullMQ queue configurations and worker code from a simple text description.
Featured*
