Abstract
Modern AI coding systems are predominantly reactive, prompt-conditioned, session-bound, and human-supervised. They generate outputs in response to explicit requests but do not persistently evolve software systems over time without user initiation. This paper introduces the Ghost Developer Protocol (GDP), a conceptual and architectural framework in which software repositories contain persistent autonomous engineering entities that continuously observe, simulate, and modify systems without external prompts. These entities, referred to as Ghost Developers, form a continuous intent layer embedded within the repository, enabling long-horizon architectural evolution, predictive refactoring, and autonomous system governance. The repository transitions from a passive artifact into an active computational organism exhibiting emergent strategic behavior, long-term memory accumulation, and self-directed architectural evolution.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Manuel-Devadoss, Johny and Prasad, Deepthi, "Ghost Developer Protocol (GDP): A Persistent Intent Architecture for Self-Evolving Software Systems", Technical Disclosure Commons, (May 12, 2026)
https://www.tdcommons.org/dpubs_series/10078