context-loop-engineering: data, code and paper for "Context Engineering and Loop Engineering for LLM Agents"
Year: 2026
Type: software
Abstract: Research compendium for the preprint "Context Engineering and Loop Engineering for LLM Agents: A Systematic Review, Unified Taxonomy, and Use-Case Fit Framework". It holds a multivocal review of 1,395 sources (544 academic, 274 grey, 577 repositories) sampled from five curated reading lists; a taxonomy of 7 context operations by 8 loop levels with a rule-based codebook audited against blind manual coding (held-out kappa 0.86); a formal model treating context as state and the loop as control, with two propositions and six Monte Carlo experiments; term diffusion over 6,278 GitHub repositories; a re-analysis of the published LoopArena results; and a use-case fit framework. Every number, table and figure in the paper is regenerated from the stored raw data with make all. Code: MIT. Paper text, figures and processed data: CC BY 4.0. Raw third-party files remain under their owners' terms.
Author: Chongmankhong, Thanatchon
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7227565493
