Dense and lean with no basic-concept padding; it assumes Three.js/engine competence and every line gives directive guidance. Recurring item lists shift context (forbid -> assert -> validate) so they earn their tokens.

Highly concrete for an instruction-only skill: explicit forbids, a literal source-to-light inventory schema (emitter ID, light ID, attachment transform, range, color/intensity, occlusion intent, fallback), and specific assertions to check.

Provides a numbered 1-6 layout sequence, a dedicated validation section (Assert/Reject/Check/Validate), and a browser-verification section with a feedback loop ('Report new failures separately from existing baseline issues').

No bundle files exist; the single document is well-sectioned into cohesive topics with no nested references, satisfying the simple-skill allowance for well-organized single-file structure.

Opens with the concrete action 'Author or revise readable, flat-world Three.js game levels' and enumerates many specific capability areas (movement/camera routes, collision/navigation, encounter zones, pickups, motivated lighting).

Explicitly answers both 'what' (author/revise flat-world Three.js levels) and 'when' via an explicit 'Use for ...' trigger clause.

The 'Use for' clause covers natural terms users would say ('Three.js game levels', 'collision', 'objectives', 'pickups', 'landmarks', 'lighting', 'visibility'), with only minor jargon ('deterministic level data', 'motivated lighting').

'flat-world Three.js game levels' carves a clear, narrow niche with distinct triggers unlikely to fire for unrelated skills.