Conciseness

The body is lean and directive throughout — "Não pesquisar novamente. Lacuna vira `claimGaps`, remoção ou reformulação." — with zero padding or explanation of concepts Claude already knows; the "Evidência de Conclusão" checklist consolidates requirements rather than duplicating them. Every section earns its tokens, matching the anchor-5 example.

Actionability

Concrete, specific guidance for an instruction-only skill: a fully specified output contract ("schemaVersion: copy-deck.v1", "três ou mais `candidates[]` com `id`, ângulo, headline, hook, corpo/roteiro, CTA"), explicit prohibitions ("Nunca invente preço, desempenho, comparação, garantia, certificação, depoimento...") and channel/duration budgets. It stops short of anchor 5 because no example JSON or worked candidate is shown, and gate execution details are deferred to external policy files.

Workflow Clarity

There is an explicit validation feedback loop — "item que falhar volta pra reescrita naquele trecho especificamente, roda o gate de novo; não entregar com item pendente" — plus a completion checklist and a failure-path handoff ("retorna `claimGaps` e volta à skill 70"). It is below anchor 5 because the end-to-end order of operations is distributed across sections rather than presented as a sequenced workflow, and the gates' actual procedures live outside this file.

Progressive Disclosure

The skill has no bundle files; the ~80-line body is compact and well-sectioned, and its external governance references (`GLOBAL.md`, `policies/anti-ai-writing.md`, `policies/source-driven.md`, `policies/evals.md`, `policies/handoffs.md`) are one level deep and clearly signaled in a single "Governança Global" section with no nesting. Per the simple-skill scoring note, well-organized sections with no need for bundle references merit the top score.

Specificity

"Transforma estrategia de campanha em mensagens e copy" names the domain and a single concrete transformation action, but does not enumerate the skill's concrete outputs (candidate routes, headlines, hooks, CTAs, channels) the way the anchor-4/5 examples list multiple actions. It is above anchor 2 because the action stated is concrete, not generic.

Completeness

Both parts are present and the "when" is explicit via the "Trigger em:" clause with concrete phrases, so the anchor-3 cap for missing trigger guidance does not apply. However, the "what" is a single terse clause that omits the deliverable (CopyDeckArtifact, distinct candidate routes) and is less comprehensive than the anchor-5 example's multi-action statement, matching anchor 4 where one half could be more specific.

Trigger Term Quality

Ten trigger phrases with natural invocations and deliberate accent variants ("variações de anúncio com evidência"/"variacoes de anuncio", "transforma a estratégia"/"transforma a estrategia") give good keyword coverage. It falls short of anchor 5 because a few natural phrasings a user might say (e.g., plain "copy de anúncio", "headline", "copywriting de campanha") are absent.

Distinctiveness Conflict Risk

Highly distinctive niche triggers like "copy com claim ledger", "copy rastreável" and "copy deck da campanha" create a clear niche with minimal conflict risk, but broader phrases such as "variacoes de anuncio" or "transforma a estrategia" could overlap with the pipeline's other copy-related skills (50, 13), which the description does not disambiguate. This places it at anchor 4 rather than 5.