Most tools are cleanly separated by purpose: search vs read vs read_many vs project_diff are unambiguous. The one soft overlap is project_map vs project_overview, which both return a directory tree, though the descriptions make the contents-vs-no-contents distinction explicit. project_dependencies is clearly distinct in output shape.

All names are snake_case, but the convention is applied unevenly: some carry a project_ prefix (project_map, project_overview, project_dependencies, project_diff) while the core file operations are bare verbs (search, read, read_many). Readable, but an agent can't infer the prefix rule predictably.

Seven tools is well within the ideal range and each one earns its place: three discovery tools, two read tools, a dependency tool, and a diff tool. No redundant or filler entries.

The read-only exploration lifecycle is largely covered: map, bulk overview, search, single/bulk read, dependency graph, and git diff. Minor gaps remain, such as no line-range or offset read for large files and no pattern/glob filtering when listing files, all of which agents can work around with search.