openclaw-hardened

Jarvis: The Agent That Doesn't Forget — complete memory and meta-learning architecture for AI agents. Three-tier memory with trust scoring and decay, nine meta-learning loops, prompt injection defense, autonomous heartbeat cycles, and the Claw Score audit rubric. Use when setting up a new OpenClaw agent, upgrading an existing agent's architecture, or when the user wants a fully autonomous, self-improving operator.

X/Twitter CLI for reading, searching, and posting via browser cookies. Use when the user asks about tweets, X/Twitter content, or wants to post/reply on X.