LLM Digest

Story

arxiv_cs_cl · Aug 25, 2026 · paper

Source brief

Recursive Experiential-Working Memory Evolution for Long-Horizon Agent Harnesses

arxiv.orgAug 25, 2026
original source linked

In brief

Recursive self-improvement (RSI) remains hard in long-horizon tasks, where growing histories obscure the task state and misalign skill invocation. We introduce Recuris, a recursive Experiential-Working Memory architec...

Feed lens

agentharness

Read the original at arxiv.org →Open in live feed Read that day’s brief

Recursive Experiential-Working Memory Evolution for Long-Horizon Agent Harnesses

Earlier in this thread 4 items

StarHarness: Evolving Harnesses with Stratified Search for Enterprise Environments

The Evolution of the Agent Harness

Recursive Self-Improvement

smevals - a small eval suite for evaluating models, prompts, and harnesses