Code Navigation & File Reading

To find symbol definitions in a large codebase, use definition-first exploration combined with targeted reading. This approach locates exact symbols, traces their callers, and queries structural dependencies to keep your context lean while navigating unfamiliar repositories.

The best way to assess dependency impact is by performing blast-radius dependency checks before making changes. This structural query identifies all callers and dependencies of a target symbol, ensuring you understand the full scope of potential edits across the repository.

Call graph analysis helps navigate unfamiliar repositories by discovering entry points and tracing neighborhoods of call chains. This reveals the underlying architecture beyond immediate references, showing how different components interact at a structural level.

To avoid reading irrelevant sections when exploring large codebases, use outline-first exploration with targeted line-range reads. By searching file outlines and requesting exact symbol definitions before opening full files, you keep context lean and focus only on relevant code.

Yes, targeted reading applies to non-code files like config and JSON. You can use config and JSON-oriented readers alongside targeted Markdown scanning to extract specific settings without performing unnecessary full file reads, maintaining efficient context usage.