Conciseness

The body assumes Claude's competence throughout — it never explains what a map, projection, or library is — and every step is decision guidance rather than padding (e.g. "screen-stable strokes for borders, outlines, and graticules unless changing thickness is itself meaningful"). It is not level 5 because the long enumerations in step 13 ("D3 geo for custom projections, thematic cartography, and annotation-rich analytical maps" and eight more) and step 9 restate tool capabilities in near-prose that could be tightened or delegated to the stack-selection references.

Actionability

As an instruction-only skill it has no code, which the scoring notes say not to penalize when guidance is actionable — and it is: concrete decision rules like "audit coordinate alignment across basemap/texture, event markers, labels, hit testing, camera focus, and fallback geometry before coding. Record longitude origin, wrap policy, and at least a few known-place spot checks" and situation-specific overlap strategies ("exact-location anchor dots when large proportional... point symbols must preserve the true coordinate"). Not level 5 because the common implementation cases (e.g. an actual choropleth recipe) are delegated to references rather than shown, leaving minor gaps in copy-paste-ready guidance.

Workflow Clarity

The 18 steps form a coherent sequence from the map-or-not decision through purpose, basemap, map family, overlap, zoom, projection, and finally stack choice, with an Output Expectations checklist. Validation checkpoints are present (step 12's pre-coding "audit coordinate alignment... before coding" and step 17's "verify the interaction with mouse wheel, trackpad, touch pan, and pinch") but there are no explicit feedback loops (validate → fix → retry), which keeps it at the 'most checkpoints present, minor validation gaps' anchor rather than 5.

Progressive Disclosure

Structure is good: an overview body, a References section split into shared theory and skill references, and all 9 skill-local reference files (e.g. `./references/map-or-not.md`, `./references/projections-and-normalization.md`) exist and match their cited paths one level deep. It falls short of the anchor-5 'well-signaled, easy navigation' pattern because the 5 shared-theory paths (e.g. `../../references/foundations/mobile-first-responsive-visualization.md`) do not resolve anywhere in the provided bundle — more than a third of the cited references are unreachable as written.

Specificity

The description lists multiple concrete actions — "deciding whether to use a map", "choosing projections or basemaps", "building choropleths or symbol maps", "implementing thematic maps, slippy maps, or geospatial interactions" — with no vague filler. It clearly matches the comprehensive-coverage anchor and not level 4, since there are no meaningful gaps in the action inventory; it is not below because nothing is generic.

Completeness

It explicitly answers both questions: "Design geospatial and cartographic visualizations" (what) and "Use when the user needs help deciding whether to use a map, choosing projections or basemaps..." (when, with concrete trigger phrases). This is the anchor-5 pattern; level 4 would require the 'when' clause to be less explicit, which it is not.

Trigger Term Quality

Natural user phrasing is covered densely: "map", "projections", "basemaps", "choropleths", "thematic maps", "slippy maps", plus thirteen named tools (Leaflet, MapLibre, Mapbox GL JS, Google Maps, OpenLayers, deck.gl, CesiumJS, etc.) that users name verbatim. This matches the comprehensive-synonyms anchor; level 4 would require noticeably missing common terms, and none are apparent for this domain.

Distinctiveness Conflict Risk

The mapping/cartography niche is distinct with map-specific triggers (choropleth, basemap, projection, slippy map) unlikely to fire for non-map skills. Minor overlap risk remains with a sibling general chart/dataviz skill, since users asking to "visualize this dataset" could plausibly trigger either — matching the 'mostly distinct, minor overlap with closely related skills' anchor rather than the minimal-conflict anchor at 5.