🛠️ Zeroshot - Executor-Verifier Orchestration for Code

Zeroshot tackles the reliability problem in AI-generated code by implementing a strict executor-verifier orchestration pattern. Instead of relying on a single LLM prompt to both write and review software changes, this tool separates the generation of code from the verification process, creating a more robust agentic workflow. Written in Rust and operating primarily as a CLI, it integrates with models like Claude and Gemini. The interesting part is not the model integration itself, but the architectural choice to enforce an independent verification loop. Many autonomous coding agents struggle with introducing regressions; Zeroshot attempts to mitigate this by formalizing the review step within the agent loop. This is a developer tool aimed at those who want to automate software changes with a higher degree of confidence. It serves as a solid reference for building orchestration engines that prioritize code safety over raw generation speed.