Scored across 44 tools
The set is deliberately split into explore_* (rendered/interactive) and ephemeris_*/vedic_*/human_design_* (raw JSON) twins, and the descriptions do state this, but several pairs (explore_natal_chart vs ephemeris_natal_chart, explore_moon_phase vs ephemeris_moon_phase, explore_vedic_chart vs vedic_chart, explore_human_design vs human_design_chart, explore_transit_timeline vs ephemeris_transits) are near-duplicates that an agent can easily misselect. Overlap also exists between electional_moment_analysis, explore_moon_phase and ephemeris_moon_phase for 'what's the sky like now'. The find_calculation → dev_read_api/dev_write_api chain, by contrast, is well delineated.
All names are snake_case and mostly verb_noun, but the prefix scheme is applied unevenly: DV prefixes (ephemeris_, explore_, human_design_, vedic_, bazi_, acg_, auth_, dev_) mix domain, transport and 'explore' semantics, and the same chart concept appears as human_design_chart, vedic_chart and ephemeris_natal_chart. The pattern is readable but not predictable from name alone.
44 tools far exceeds a scannable surface, and a meaningful share is duplicated UI/data variants of the same calculation. The breadth of the astrology domain justifies many capabilities, but the set would be better served by fewer tools with a format/mode parameter than by per-variant endpoints.
Coverage is unusually deep: natal, transits, progressions, returns, synastry, composite/davison, relocation, Vedic with dashas, BaZi, Human Design (natal, transit, connection), astrocartography, electional, time lords, plus auth/billing and a catalog-driven escape hatch. Gaps are minor: composite/davison and a few advanced techniques exist only in rendered form, and there is no persistence layer for saved charts.