Ripple - Predict Code Change Impact

Changing one file can unexpectedly affect functions, modules, dependencies, tests, and other parts of a software project. Developers often manually trace references, inspect imports, understand call relationships, and determine which tests may need rerunning. This slows development and increases the risk of missed dependencies and regressions. Ripple is a developer workflow that helps teams understand the potential consequences of a code change before implementation. A developer connects or uploads a repository, and Ripple scans it to build a code intelligence graph containing imports, function calls, references, modules, classes, dependencies, and tests. Developers can simulate proposed changes, including modifying, deleting, renaming, or adding files, functions, or modules. Ripple traverses the repository graph to identify the potential blast radius and organizes evidence into direct impact, indirect impact, potentially broken areas, and tests to rerun. This gives developers a structured view of what may be affected before a change and what should be verified afterward. Ripple integrates IBM Bob 2.0 through the Model Context Protocol (MCP). Bob provides the AI development interface, while Ripple provides specialized repository intelligence. Through the Ripple MCP server, Bob can access Ripple's analysis capabilities and use repository-grounded results to explain the consequences of proposed changes. Ripple is designed for software developers, engineering teams, code reviewers, and teams working with unfamiliar or complex repositories. Its key distinction is the separation between AI reasoning and code intelligence: Ripple does not rely on an LLM to guess the blast radius. Its core analysis is grounded in actual repository relationships and graph traversal, while Bob provides an interactive AI interface for understanding those results. IBM Bob 2.0 is a core component of both Ripple's development process and its developer-facing workflow. We used Bob throughout development to design, implement, debug, refine, and iterate on Ripple's architecture, frontend, backend, UI, APIs, repository analysis, and features. More importantly, Ripple exposes an MCP server that IBM Bob can connect to. In our development environment, Bob connects to Ripple using a Streamable HTTP MCP connection. This enables Bob to interact with Ripple's specialized repository-analysis capabilities rather than relying only on general coding knowledge. The workflow is: IBM Bob → Ripple MCP → Ripple Code Intelligence → Structured Analysis → Bob. When a developer asks Bob about the potential impact of a proposed code change, Bob can use Ripple's MCP capabilities to obtain repository-grounded information. Ripple analyzes the repository's code graph and returns structured information about affected relationships and potential consequences. Bob can then reason over those results and explain them to the developer. This creates a clear separation of responsibilities: Bob provides the AI interface and reasoning layer, while Ripple provides specialized, repository-grounded code intelligence and blast-radius analysis. Ripple is not simply a chatbot wrapped around an LLM; its core impact analysis is based on the structural relationships within the connected repository. IBM Bob acts as both a development partner and an intelligent interface for Ripple through MCP. IBM Bob task-session screenshots are included in the project's bob_sessions directory as submission evidence.