Hippocampus Subagent Memory provides isolated memory namespaces for sub-agents to prevent context contamination and manage explicit data flows.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install hippocampus-subagent-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 hippocampus-subagent-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 Hippocampus Subagent Memory?
Hippocampus Subagent Memory is a specialized tool within the Openclaw Skills ecosystem designed to solve the challenge of memory leakage when spawning sub-agents. By implementing scoped IDs and bounded merge-back logic, it ensures that every sub-agent operates within its own secure namespace, maintaining the integrity of the parent agent context.
This skill is essential for complex multi-agent architectures where data privacy and context precision are paramount. It allows developers to build robust Openclaw Skills that can scale through hierarchical agent tasking without the overhead of manually managing transcript noise or conflicting memory states.
Hippocampus Subagent Memory Use Cases
- Creating isolated sub-agent memory namespaces to prevent sibling data leakage.
- Returning only bounded summaries or specific artifacts to the parent agent.
- Coordinating explicit imports from child memory back to the parent context.
- Preventing accidental context contamination during multi-step sub-agent tasks.
How Hippocampus Subagent Memory Works
- The parent agent spawns a child agent using a unique, scoped HIPOKAMP_SUBAGENT_ID.
- The child agent restricts all memory writes strictly to its designated isolated namespace.
- Upon task completion, the child agent prepares a bounded result package containing only essential data.
- The parent agent explicitly approves and imports specific summaries or references from the child package.
Hippocampus Subagent Memory Setup
To integrate this skill into your Openclaw Skills workflow, configure your environment with the following steps:
# Install the necessary dependencies
npm install @hippocampus/openclaw-context-engine
# Set the sub-agent ID environment variable
export HIPOKAMP_SUBAGENT_ID="agent-task-unique-id"
Ensure that the hippocampus-openclaw-onboarding process is completed before initializing sub-agent memory namespaces.
Hippocampus Subagent Memory Data Schema & Taxonomy
The skill organizes data through a hierarchical namespace structure using session and linkage metadata.
| Metadata Key | Description | Scope |
|---|---|---|
| HIPOKAMP_SUBAGENT_ID | Unique identifier for the sub-agent session | Local |
| parent_linkage | Reference ID linking the child to its parent agent | Global |
| memory_type | Categorization of data (Summary, Artifact, Reference) | Local |
| session_id | The root session ID shared across the agent hierarchy | Global |
Hippocampus Subagent Memory Advanced Features
- Automated spawn and end lifecycle hooks via the @hippocampus/openclaw-context-engine integration.
- Metadata-based tagging for detailed session and parent linkage tracking.
- Bounded merge-back logic to ensure only high-value data enters the parent context.
- Cross-agent import controls for granular data governance within your Openclaw Skills.
SKILL.md
Loading
Related Openclaw Skills
A-Share Ultra-Short Trading Strategy
A professional-grade ultra-short trading framework designed for A-share market leaders and high-volatility T+1 arbitrage.
Argentina
An expert AI agent skill for planning complex Argentina travel itineraries with a focus on local logistics, money strategy, and regional route optimization.
RDK X5 Quickstart Guide
An essential onboarding guide for the RDK X5 platform that leads developers from hardware setup to running real-time AI object detection.
NewsNow Reader
A lightweight Python-based news aggregator that fetches real-time trending topics from eight major Chinese media platforms without external dependencies.
Small-Goods Loyalty & Incentives (DTC)
A specialized tool for designing high-frequency loyalty programs and repeat purchase incentives for DTC brands selling small-ticket items.
Reinforced Thinking Mode
A high-intensity cognitive framework for AI agents to perform multi-round, independent, and divergent deep thinking for complex architectural and strategic tasks.
Featured*
