Most tools map to distinct steps in a linear exam-figure workflow (guide, data lookup, region selection, value prep, labeling, figure generation), and descriptions clarify ordering such as calling guide first and list_bundled_data before kosis_search. The data-discovery cluster (list_sources, list_bundled_data, kosis_search, kosis_items) has some conceptual overlap that could cause initial hesitation, but each has a differentiated scope.

All names use consistent snake_case, and most follow a verb_noun pattern (list_*, find_*, set_*, prepare_*, assign_*, make_*, suggest_*). Minor deviations exist: 'guide' is a bare noun and 'kosis_items'/'kosis_search' put the resource noun first, but the set remains predictable and readable.

At 14 tools, the server sits comfortably within a well-scoped range and each tool corresponds to a plausible workflow step. The surface is slightly heavy in the data-discovery/listing area and includes two configuration helpers (set_kosis_key, set_style), but nothing is clearly redundant.

The set covers the full pipeline from source discovery and statistics retrieval through region suggestion, value transformation, label assignment, and PPTX/HWPX figure generation. Minor gaps exist around managing outputs and settings—no tool to read/reset saved style, list or inspect generated files, or revise an existing figure spec—but core exam-figure workflows are covered.