foxctl SWE Grep

You can extract code snippets by providing a workspace ID, a natural-language question, and candidate file paths to retrieve high-signal excerpts. The tool reads the live workspace and selects the most relevant lines matching your query.

Evidence-aware code retrieval is a workflow where semantic or symbol indexes first shortlist likely files, then live workspace reading extracts the exact snippets. You need this approach when searching large codebases to answer how specific functions or handlers work.

Yes, you can enforce configurable limits on max files, max snippets, and max bytes per file during code retrieval. These precision controls keep results relevant and efficient while prioritizing candidate files based on assigned priority levels.

Symbol hints improve code snippet extraction by providing optional symbol IDs alongside candidate paths. This guides the retrieval process to focus on specific functions or variables, ensuring extracted excerpts better answer your natural-language question.

Semantic search alone identifies candidate files but cannot extract exact code snippets. You need a final retrieval stage that reads the live workspace to pull the necessary lines, using optional symbol hints and priority levels for focused results.

The best way to locate existing code patterns is running snippet extraction with a natural-language question and candidate file paths. This pulls high-signal excerpts from the live workspace, revealing current patterns to guide feature planning and modifications.