deep-research

Comprehensive code understanding of unfamiliar modules is achieved through a structured, multi-phase research protocol. It leverages memory-first analysis, comprehensive LSP exploration, and pattern discovery using grep and glob to map interactions before implementation.

The best way to explore code interactions across multiple files is via an explore agent utilizing comprehensive LSP operations. This method systematically maps dependencies and structural relationships when dealing with complex features.

Yes, grep and glob can be used for pattern discovery when researching complex codebases. They are integrated into the research protocol to identify existing solutions and patterns after initial memory-first analysis.

Researching a complex authentication middleware implementation involves a multi-phase protocol: memory-first analysis, LSP exploration, pattern discovery, and targeted external research. This ensures thorough code comprehension prior to editing.

External research should be used during codebase exploration when internal analysis methods are insufficient. It is a targeted phase applied after memory-first analysis, LSP exploration, and pattern discovery fail to provide needed context.

Yes, deep code research works without external dependencies. It relies on internal mechanisms like memory search, LSP operations, and pattern discovery, utilizing targeted external research only when internal methods prove insufficient.