Conciseness

The body is ~25 lines of pure instruction with zero padding: 'Respond with short, direct language', six terse rules, and a three-level intensity scale. It explains nothing Claude already knows and every token earns its place — a direct match for anchor 5.

Actionability

Guidance is concrete and directive for an instruction-only skill ('Prefer fragments over long prose', 'Omit praise, hedging, filler, and recap unless needed', 'briefly explain the risk before proceeding'), with an explicit default level. It stops short of anchor 5 because the three intensity levels are named but never exemplified — no sample output or before/after showing what Level 2 vs Level 3 responses look like.

Workflow Clarity

This is a simple, single-purpose skill where the rubric's simple-skill exception applies: the single action (apply the rules when responding) is unambiguous. It also embeds a validation-style checkpoint — 'For dangerous, destructive, expensive, or ambiguous work, briefly explain the risk before proceeding' — and explicit default-level handling ('Default to Level 2 unless the user names another level').

Progressive Disclosure

The skill is under 50 lines with no need for external references (no references/, scripts/, or assets/ files exist in the bundle), and content is organized into clean 'Rules' and 'Intensity' sections — exactly the case the rubric's simple-skill scoring note says can score 5.

Specificity

The description states several concrete behaviors — 'Keep technical accuracy, preserve necessary safety or ambiguity warnings' and an explicit exclusion list ('do not apply this style to code, commits, PR descriptions, or user-facing product copy') — with only minor gaps (e.g., no mention of response-length scope). It falls short of anchor 5 because the core capability is a single style action framed by constraints rather than multiple comprehensive concrete actions.

Completeness

It explicitly answers both: what ('extremely terse, blunt, token-efficient communication' with accuracy and warning-preservation constraints) and when ('Use when the user asks for Caveman-style, extremely terse, blunt, token-efficient communication'), with concrete trigger phrases directly in the when-clause. This matches anchor 5; anchor 4 would require the when to be less explicit, which it is not.

Trigger Term Quality

Natural phrases a user would actually say are present — 'Caveman-style', 'extremely terse', 'blunt', 'token-efficient'. A few common synonyms users might use ('brief', 'concise', 'short answers', 'one-liners') are missing, so it matches anchor 4 rather than the comprehensive coverage of anchor 5.

Distinctiveness Conflict Risk

A clear niche — one named communication style — with an explicit boundary excluding code, commits, PR descriptions, and product copy, minimizing overlap with general communication or writing skills. Matches anchor 5: distinct triggers, minimal conflict risk.