// readme

llmXive — automated scientific discovery, conducted in the open

llmXive is a research platform that develops ideas into executable studies, reviews their evidence, and drafts papers for review and publication approval. Specialist LLM agents work through project-specific specifications, plans and tasks. Git records the artifacts, review decisions and recovery history.

Current status — 2026-10-09

The recovery audit found failures in task scope, execution, artifact placement, review protocols and evidence verification. Updating the model alone did not resolve them. The implementation now includes fixes for those observed failures, but a fresh run through the entire authored-paper pipeline is not yet proven.

An earlier isolated totient study…