Conciseness

The body is itself ultra-compressed ("Respond terse like smart caveman. All technical substance stay. Only fluff die.") and models the style it teaches; every section adds rules or examples Claude could not infer, with no padding or explanation of known concepts. Fits the 'lean and efficient; every token earns its place' anchor; not 4 because there are no instances of over-explanation to trim.

Actionability

Concrete, executable guidance: enumerated drop lists ("articles (a/an/the), filler (just/really/basically/actually/simply), pleasantries (sure/certainly/of course)"), abbreviations, the "[thing] [action] [reason]" pattern, a Not/Yes contrast, and two Q&A examples. Not 5 because the Tooling section ("Tools: compressor + estimator + lint. Agent: cs-caveman-mode. Command: /cs:caveman.") names tools with no usage instructions and points to a file absent from the bundle — a minor gap.

Workflow Clarity

A simple, single-purpose skill (under 50 lines) whose single action is unambiguous: apply the compression rules every response, with persistence rules ("ACTIVE EVERY RESPONSE once triggered... Off only when user says 'stop caveman'") and clearly sequenced exception handling ("Drop caveman temporarily for: security warnings, irreversible action confirmations... Resume caveman after clear part done") including a verification cue ("Verify backup exist first"). Per the simple-skill exception this scores 5; the destructive/batch validation cap does not apply since the skill performs no destructive operations itself.

Progressive Disclosure

Sections are well organized (Persistence, Rules, Examples, Auto-Clarity Exception, Tooling), but the sole external reference "See [references/companion_tooling.md](references/companion_tooling.md)" is dangling — no references/, scripts/, or assets/ directories exist in the bundle, so the promised compressor/estimator/lint detail is inaccessible. Not 4 because a broken referenced path is more than a minor organization gap when scored against the actual bundle structure; not 2 because the core content is well-structured and nothing that belongs in a separate file is inlined.

Specificity

"Cuts token usage ~75% by dropping filler, articles, and pleasantries while keeping full technical accuracy" lists several specific concrete actions with a clear domain (ultra-compressed communication). Not 5 because it is a single-mode transformation and coverage of the behavior is not fully comprehensive (e.g., fragment style and persistence behavior are not mentioned); not 3 because it goes well beyond naming the domain with 1-2 actions.

Completeness

Explicitly answers both: what ("Ultra-compressed communication mode. Cuts token usage ~75% by dropping filler, articles, and pleasantries while keeping full technical accuracy") and when ("Use when user says 'caveman mode', ... or invokes /caveman") with concrete trigger phrases. Matches the anchor-5 example structure exactly; the 'when' clause is explicit, not weakly implied.

Trigger Term Quality

Triggers "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", and "/caveman" are natural phrases a user would say. Not 5 because common synonyms like "terse", "concise", or "short answers" are missing; not 3 because coverage is good, not merely some relevant keywords.

Distinctiveness Conflict Risk

"Caveman mode" triggers give this a clear niche with minimal conflict risk, but the generic triggers "less tokens" and "be brief" could plausibly fire for other brevity/token-saving skills. Not 5 due to this minor overlap risk; not 3 because the skill is mostly distinct and the core triggers are unmistakable.