rgx is a candidate-index layer in front of ripgrep: the index narrows which files get scanned,
ripgrep does the matching. Correctness is ripgrep's, speed is ours.
README.md— what it is, install & use.docs/design.md— mission, model, correctness contract, open questions.docs/cli.md— command surface and the--servergate.docs/mcp.md— agent-facing MCP tools.docs/indexing.md— write path: building, storing, and keeping the trigram index fresh, plus the ripgrep-parity walk.docs/querying.md— read path: regex→trigram query, candidate selection, ripgrep confirm, output/paging.docs/profiling.md— profiling the build/query paths (criterion, samply, dhat).
Development process — build/test/lint commands, conventions, testing, and the release flow — lives in
CONTRIBUTING.md, imported below.
@CONTRIBUTING.md