Boundary relies entirely on your IDE's code intelligence to provide comprehensive context. Cursor (or similar IDEs) acts as Boundary's eyes and ears, using LSP, "Go to Definition", and "Find References" to gather the code context needed for a thorough debate.
When you initiate a debate, Cursor's AI agent becomes the semantic intelligence proxy for Boundary. Here's how they work together:
You ask Cursor to start a debate
You present your engineering decision question to Cursor's AI agent, which then calls Boundary's MCP server to initiate the debate session.
Boundary generates semantic queries
Boundary analyzes your question and codebase context, then generates specific queries about symbols, boundaries, call graphs, and implementations it needs to understand.
Cursor investigates using code intelligence
Cursor uses its superior code intelligence: LSP, "Go to Definition", "Find References", codebase search to gather comprehensive context. It extracts code snippets, traces dependencies, and finds related patterns.
Iterative context gathering
This process repeats over multiple rounds. Boundary may ask for deeper analysis of specific components, edge cases, or dependencies. Cursor continues investigating and providing context until Boundary has everything needed.
Boundary completes the debate
Once Boundary has sufficient context, it completes the multi-agent debate and returns a decision map. Cursor displays this to you, showing the risk surfaces, trade-offs, and failure modes surfaced by the debate.
Note: This diagram shows the iterative collaboration between Cursor and Boundary. The debate process requires multiple rounds of context gathering to ensure comprehensive analysis.