A professional integration skill connecting OpenClaw to Aionis for advanced long-term memory, context-aware tool routing, and feedback-driven policy loops.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install aionis
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 aionis 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 Aionis Memory Policy Loop?
The Aionis Memory Policy Loop is a sophisticated skill designed to provide Openclaw Skills with deep, persistent memory capabilities. It bridges the gap between ephemeral agent interactions and long-term knowledge retention by utilizing the Aionis API for writing facts, assembling layered context, and managing tool selection policies. This skill is essential for developers who need their agents to recall specific user preferences, project-specific rules, and historical decisions.
By implementing this loop, agents can perform retrieval-augmented generation (RAG) with high-fidelity citations and adapt their behavior over time through a dedicated feedback mechanism. It transforms a standard AI agent into a context-aware system capable of complex planning and self-improvement across multiple sessions.
Aionis Memory Policy Loop Use Cases
- Implementing long-term memory for AI coding agents to remember architectural decisions across different sessions.
- Creating retrieval-augmented generation (RAG) pipelines that include specific citations and layered context for high accuracy.
- Automating tool routing based on memory-driven policies to ensure the most efficient resource utilization.
- Establishing a feedback-driven learning loop where the agent improves its tool selection based on positive or negative outcomes.
- Managing project-specific rules and constraints that need to be injected into the agent's context during planning.
How Aionis Memory Policy Loop Works
- Connectivity Validation: The skill performs a precheck to ensure the Aionis standalone server is reachable and authorized.
- Fact Ingestion: Key facts and event results are stored as nodes in a structured memory lane using the write API.
- Context Assembly: Before planning a task, the skill builds a multi-layered context—including facts, rules, and previous decisions—tailored to the specific query.
- Policy-Based Routing: The system selects the appropriate tools by evaluating candidates against existing memory rules and historical context.
- Execution Feedback: After a tool is used, the outcome is fed back into the system to refine future routing policies and close the memory loop.
Aionis Memory Policy Loop Setup
To enable this skill within your Openclaw Skills environment, configure the following variables:
- Set the
AIONIS_BASE_URLto your local or container host (e.g.,http://127.0.0.1:3001). - Provide authentication via
AIONIS_API_KEYorAIONIS_AUTH_BEARER. - (Optional) Define
AIONIS_TENANT_IDandAIONIS_SCOPE_PREFIXfor multi-tenant isolation.
If the local standalone server is not running, use the bootstrap command:
bash ./bootstrap-local-standalone.sh
source ./.runtime/clawbot.env
Aionis Memory Policy Loop Data Schema & Taxonomy
The skill organizes data into a structured taxonomy of nodes and edges, allowing for complex relationship mapping. This is critical for Openclaw Skills to maintain high-quality context.
| Component | Description | Key Fields |
|---|---|---|
| Memory Node | Atomic data point representing an event or fact. | client_id, type, text_summary |
| Context Layers | Configurable levels of information retrieval. | facts, rules, decisions, citations |
| Feedback Object | Data used to adapt tool selection policies. | run_id, outcome, selected_tool |
| Scope | The project-specific identifier for memory isolation. | clawbot:<project_name> |
Aionis Memory Policy Loop Advanced Features
- Layered context assembly with configurable character budgets to prevent context window overflow.
- Automated embedding generation for all ingested text data to support semantic search.
- Neighborhood hop traversal for deep graph-based relationship discovery (up to 2 hops and 50 nodes).
- Shadow policy support for testing new routing rules without impacting production workflows.
- Fallback mechanisms that ensure continuous operation even if advanced assembly endpoints are temporarily unavailable.
SKILL.md
Loading
Related Openclaw Skills
Crawl4ai
An advanced AI-powered web scraping framework designed to extract structured, clean data from complex websites and dynamic JavaScript-heavy environments.
XVFB Chrome Automation
A specialized toolset for running headful or headless Chrome on Linux servers using XVFB to enable seamless AI-driven browser automation.
browsex
A browser-based CLI skill for interacting with X/Twitter through the OpenClaw agent, supporting both authenticated and read-only actions.
BrainRepo
A comprehensive personal knowledge repository that utilizes PARA and Zettelkasten methodologies to capture, organize, and retrieve information via AI agents.
Himalaya Email CLI
A fast and minimal CLI tool to manage emails via IMAP, SMTP, and Notmuch backends directly from your terminal.
Orgo Desktop Control
A comprehensive Python SDK integration for provisioning, managing, and automating remote cloud desktops and sandboxed execution environments.
Featured*
