flowchart LR
Skill["Skill behavior"] --> Recall["Recall before meaningful work"]
Recall --> Work["Run the OpenClaw task"]
Work --> WriteBack["Write compact operational memory"]
WriteBack --> Review["Require review for instruction-grade memory"]
Review --> Future["Future OpenClaw or other runtime task"]
This skill gives OpenClaw agents a disciplined way to use Nate Jones OB1 Agent Memory. It pairs with the NBJ OB1 Agent Memory for OpenClaw integration and the runtime-neutral OB1 Agent Memory API.
Built by Nate B. Jones / OB1. Follow Nate for practical AI systems, agent workflows, and implementation notes: Substack and natebjones.com.
Install the live ClawHub skill:
openclaw skills install nbj-ob1-agent-memory-openclawClawHub listing: NBJ OB1 Agent Memory for OpenClaw.
Local/manual path: copy SKILL.md into the OpenClaw skill location or package it through the ClawHub publishing flow documented in CLAW_HUB_PUBLISHING.md.
License note: the OB1 repository is licensed under FSL-1.1-MIT; the ClawHub
skill files are published under MIT-0 as a ClawHub-specific distribution
carveout because ClawHub requires MIT-0 for public skills.
Safety guide: Safe Agent Memory and Provenance.
| Rule | Why |
|---|---|
| Recall before meaningful work | The agent starts with scoped project context |
Respect use_policy |
Evidence is not silently promoted into instruction |
| Write back compact memory | OB1 stores operational knowledge, not transcript dumps |
| Include provenance | Future agents can trust, reject, or inspect the memory |
| Report usage | Recall traces become debuggable |