Loading…
Programmatic memory for long-horizon LLM agents: the harness appends everything to one log, and the agent searches it with code. 97.4% on ARC-AGI-3 (arXiv:2607.20064)
Our verdict: worth using
Provides a plug‑and‑play durable memory layer for Claude Code/Claude Code‑compatible agents, exactly the long‑horizon agent memory the engineer needs
Filed under agent memory in our directory.