Conciseness

The body is four lean sentences covering only what Claude would not already know — the hook delivery path (`hooks/caveman-stats.js`, `hooks/caveman-mode-tracker.js`) and the `decision: "block"` mechanism. There is no padding, no explanation of known concepts, and every token earns its place, matching anchor 5.

Actionability

The guidance is concrete and executable for the intended path — "The model does not need to do anything when this skill fires — the hook returns `decision: \"block\"` with the formatted stats as the reason" — with real file names and hook semantics. It falls short of anchor 5 only because there is no fallback instruction (e.g., what to do if the hook is not installed or fails), leaving a minor gap.

Workflow Clarity

This is a simple, single-purpose skill where the entire workflow is one unambiguous action: the model does nothing and the hook handles everything. Per the simple-skill exception, the unambiguous single action merits 5, and no destructive or batch operations exist that would trigger the validation cap.

Progressive Disclosure

The body is under 50 lines with no content that warrants external reference files; the inline pointers to `hooks/caveman-stats.js` and `hooks/caveman-mode-tracker.js` are one level deep and clearly signaled, and no references/ or scripts/ bundle exists. Per the under-50-lines guideline this matches anchor 5.

Specificity

The description lists several concrete actions and mechanisms — "Show real token usage and estimated savings for the current session", "Reads directly from the Claude Code session log — no AI estimation", and "Output is injected by the mode-tracker hook" — which goes beyond naming 1-2 actions (anchor 3) but stops short of comprehensive multi-action coverage (anchor 5).

Completeness

The "what" is clearly answered ("Show real token usage and estimated savings for the current session") and the "when" is explicit and concrete ("Triggers on /caveman-stats"), matching the anchor-5 example that answers both with concrete trigger phrases. It is not score 4 because the trigger is not merely implied or generic — it names the exact invocation.

Trigger Term Quality

Natural terms like "token usage", "estimated savings", and "session" are present alongside the explicit trigger "Triggers on /caveman-stats", giving good keyword coverage for a user who wants usage stats; a few natural synonyms (e.g., "cost", "spend", "token count") are missing, which fits anchor 4 rather than anchor 5.

Distinctiveness Conflict Risk

The combination of a specific niche (session token usage and savings reported from the session log) with a distinct, unique slash-command trigger (/caveman-stats) gives minimal conflict risk with other skills, matching anchor 5 rather than anchor 4's "minor overlap risk with closely related skills".