agentleFS
moronireismoronireis/agent-system-partners/.claude/skills/humanization/SKILL.md

humanization

Skill0 starsChanged 2 months agoMITView on GitHub

PT-BR AI-pattern elimination — banned words, banned structures, sentence-rhythm rules, and natural executive-voice injection. Use when mkt-editor runs the humanization pass on any copy or content draft before the quality-gate verdict.

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 6 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Remove on sight, no exceptions:

    line 12
  2. Distinctive:

    Use intentional fragments for impact.

    line 35
  3. Distinctive:

    No more than 2 consecutive sentences may open with the same grammatical structure.

    line 36
  4. Distinctive:

    No paragraph should run uniform sentence length throughout — check this explicitly.

    line 37
  5. Distinctive:

    Every number, data point, and proof element already present in the draft

    line 51
  6. Distinctive:

    If any answer fails, rewrite that section before moving to the quality-gate scoring step.

    line 71

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Humanization

A PT-BR voice pass that eliminates AI-generated patterns and injects natural executive register. This skill governs mkt-editor's humanization pass — it changes voice and rhythm only. It never changes strategy, structure, claims, or numbers already in the draft.

Banned Words & Expressions (PT-BR)

Remove on sight, no exceptions:

  • Hype words: "revolucionário", "inovador", "solução robusta", "cutting-edge", "disruptivo"
  • Inflated verbs: "transformar" (as a vague catch-all verb), "alavancar", "potencializar", "impulsionar"
  • Generic openers-as-phrases: "neste artigo/guia vamos explorar...", "no mundo atual de...", "com o avanço da tecnologia..."
  • Conclusion filler: "em conclusão", "em resumo", "para finalizar"
  • Hedge-padding: "é importante ressaltar que", "vale destacar que", "sem dúvida", "certamente", "indubitavelmente"
  • Abstract-noun crutches: "jornada", "ecossistema", "sinergia", "paradigma"
  • Formal connectors that read as translated: "além disso", "ademais", "outrossim", "nesse sentido"
  • "navegar" outside a literal nautical context
  • Any run of 3+ stacked adjectives

Banned Structures

  • Generic openings ("No mundo atual de...", "Com o avanço da tecnologia...")
  • Lists of 5+ items that all start with the same grammatical structure
  • Paragraphs where every sentence runs close to the same length
  • Conclusions that just repeat the introduction in synonyms
  • Transitions that announce themselves ("Dito isso,", "Nesse contexto,") instead of just connecting ideas

Sentence-Rhythm Rules

  • Vary length deliberately: a short sentence, then a longer one that actually develops the idea, then short again.
  • Use intentional fragments for impact. Like this one.
  • No more than 2 consecutive sentences may open with the same grammatical structure.
  • No paragraph should run uniform sentence length throughout — check this explicitly; it is the single most common AI tell.

Natural Executive Voice Injection

  • An occasional conversational opener ("Olha,") — used sparingly, once or twice per piece, never more.
  • Plain transitions over formal ones: "na prática" instead of "em termos práticos"; "funciona assim:" instead of "o processo se dá da seguinte maneira:".
  • Parenthetical asides that read like someone thinking out loud — used sparingly, not on every paragraph.
  • An occasional self-correction ("na verdade, melhor dizendo...") signals a real person drafting, not a polished machine.
  • Emotional specificity over abstraction: replace an abstract benefit with the concrete moment it's felt — always tied to the actual claim already in the draft, never an invented detail.
  • Contractions and rhythm natural to Brazilian oral-formal register — not academic, not slang-casual, not translated-sounding.

What Never Changes During Humanization

  • Strategic frameworks and persuasion architecture (hooks, CTAs, proof stacking, objection handling)
  • Every number, data point, and proof element already present in the draft
  • Section structure and overall flow
  • Claims — humanization is a voice pass, not a fact pass. Never invent a detail to sound more natural.

Before / After

Before (AI pattern): "Nossas soluções inovadoras permitem que empresas transformem seus processos e alcancem resultados extraordinários através de implementações robustas e escaláveis."

After (humanized): "A maioria das empresas que chegam aqui já tentou isso antes. Gastaram, não funcionou. O que muda agora? Começar pelo resultado financeiro, não pela tecnologia. Simples assim."

Final Checklist (read the piece mentally out loud)

  1. Uma pessoa real diria isso numa reunião de verdade?
  2. Soa como algo que saiu de um chatbot?
  3. O ritmo está monótono — todas as frases do mesmo tamanho?
  4. Tem 2+ frases seguidas começando igual?
  5. O texto foi escrito PARA o leitor, ou só pra preencher espaço?

If any answer fails, rewrite that section before moving to the quality-gate scoring step.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.