Work

Self-Replicating LLM Artifacts & Accidental Supply-Chain Contamination

Year: 2026

Type: article

Abstract: This paper introduces the concept of "Logical Prions": self-replicating LLM artifacts that exploit recursive logic to induce systemic failure. By documenting a real-world case of supply-chain contamination involving a shell bootstrap installer, this research explores how Ken Thompson’s "Reflections on Trusting Trust" manifests within modern stochastic systems. The author identifies and analyzes a new class of vulnerability where metaprogramming triggers recursive reasoning loops, presenting a significant risk to LLM-assisted developer workflows.

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.18582000

Author: Myint Jr, James

Language: English

Topic: Security and Verification in Computing

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

OpenAlex ID: W7128520510

FWCI:

Cites:

Cited by:

Related to: