Show HN: Open Kioku – local evidence layer for AI coding agents (github.com)

Open Kioku has launched as a powerful local evidence layer designed to enhance AI coding agents, such as Claude, Cursor, and Codex, by providing pre-edit context before making changes to a codebase. Users can install Open Kioku via npm and generate local indexes that include vital information like symbols, references, and potential impacts of proposed edits. This process enables AI agents to execute a sequence of commands—searching code, building evidence-backed plans, and ultimately verifying changes—thus streamlining their interaction with the codebase and reducing inefficiencies. The significance of Open Kioku for the AI/ML community lies in its ability to empower coding agents with a more informed decision-making process. By indexing local repositories, Open Kioku ensures that AI agents retrieve accurate and relevant information, minimizing the risk of unwanted changes and errors. Technical details show that a local repo audit can index thousands of files and symbols rapidly, while generating comprehensive reports for assessment. This innovation may transform how AI interacts with code, making it more reliable and efficient, thereby accelerating software development workflows and enhancing collaboration within development teams.