A Datalog engine for LLM agent memory: stratified rules, provenance-tracked facts, incremental derivation, and an MCP server that lets your harness use it as a shared brain.
A neurosymbolic perspective on LLMs
Coding Agents skills for Synalinks OSS
Synthesized Logic Language for World Models π±ππ
A lisp built for self-modifying AI agents