Netflix Releases Preview of 'ja', an Agent-Friendly CLI Toolset Centered on the Java Module System

Netflix's JVM ecosystem team announced a preview of 'ja', a composable, module system native, agent friendly CLI toolset for modern Java development. While the existing Java ecosystem is mature, a limitation remained where module descriptors were treated merely as part of a project's description even after the introduction of the Java Module System. 'ja' elevates the module descriptor to the complete description of the project, clearly expressing dependency versions and metadata through requires directives and documentation tags.

Agent-Friendly CLI Tool Composition

Due to the existing IDE-centric ecosystem, CLI tools were lacking, causing coding agents to struggle with identifying dependencies, documentation, and source locations. To address these issues, 'ja' provides a toolset where each feature is implemented as an independent standalone tool.

• jig: Performs module version resolution, compilation, and assembly, acting as a bridge to Maven repositories
• jfmt: Formats source code based on Java programming language conventions to prevent formatting issues in agent-generated code
• jist: Provides accessibility to agents without LSP or MCP through source-aware symbol search via a grep-style interface
• jdocserver: Provides locally navigable API documentation

Integration Strategy with the Maven Ecosystem

A survey of the top 1,000 artifacts on Maven Central revealed that only 232 had explicit module definitions and 248 had automatic module names. 'ja' leverages the verified namespaces of Maven Central (reverse-domain group IDs) to define canonical Maven module coordinates. This allows most existing artifacts to be discovered by module name and version alone, while existing module authors are encouraged to publish a single Maven relocation pom to support discovery of original coordinates.

Security and Integrity Enhancements

To address technical debt caused by classpath misuse, 'ja' declares runtime access requirements as module metadata. Libraries record the necessary access, and consuming applications must explicitly approve it. Without approval, dependency resolution fails, and persistent hashes of resolved binary dependencies are stored in module-info.hash files to ensure module integrity. Additionally, annotation processing is treated as an explicit code generation step, ensuring that generated sources are visible during code review.