agentleFS
CodeCaroSCodeCaroS/MythOs_Aegis/.agents/skills/humanizer/SKILL.md

humanizer

Skill0 starsChanged 3 months agoApache-2.0View on GitHub

Rewrite or review German or English prose so it sounds natural, specific, and consistent with the author's voice without changing facts, certainty, or protected structure. Use for documents, emails, posts, PR descriptions, comments, status updates, and agent responses that feel generic, promotional, repetitive, over-polished, or AI-written. Also use when the user asks to humanize text, remove AI tone, match a writing sample, or make wording sound more like them. Do not use for code-only tasks.

188 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use economy-mode, personal-voice and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • Microsoft AzureSkill

    economy-mode

    Shifts Layer 3 model selection to cost-optimized alternatives when economy mode is active.

    Also in 4 projects that use this

  • petermuirurithuoSkill

    personal-voice

    Captures the user's personality, communication style, and values. Use this skill whenever responding to the user, drafting anything on their behalf (emails, CVs, messages, opinions, explanations), or whenever tone, word choice, or how an opinion is delivered matters. Apply it by default for any prompt, not only when the user explicitly asks for "my style" or "my voice.

    Also in 3 projects that use this

  • Microsoft AzureSkill

    architectural-proposals

    How to write comprehensive architectural proposals that drive alignment before code is written

    Also in 3 projects that use this

  • +9 more from Azure

The file

We split this skill into the rules it gives an agent

12 rules. 0 are common, 12 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 175 edited versions we read keep these.

  • 12

    Distinctive

    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:

    Follow the user's explicit wording, tone, language, and format requirements.

    line 25
  2. Distinctive:

    Do not alter or invent:

    line 32
  3. Distinctive:

    Keep domain terms when they are precise.

    line 39
  4. Distinctive:

    Look for clusters of problems, not isolated words or punctuation:

    line 45
  5. Distinctive:

    Read the result once for voice and once for factual fidelity.

    line 58
  6. Distinctive:

    When another task invokes this skill for an email, document, PR description, commit message, or final response, apply the pass silently and …

    line 77
  7. Distinctive:

    Run factual, security, and reasoning checks before this style pass.

    line 83
  8. Distinctive:

    Do not soften criticism, hide uncertainty, or turn evidence into opinion.

    line 84
  9. Distinctive:

    Keep concise harness responses concise.

    line 85
  10. Distinctive:

    Do not add a "humanization audit" unless requested.

    line 86
  11. Distinctive:

    every factual claim still matches the source.

    line 92
  12. Distinctive:

    no protected token or required section changed accidentally.

    line 93

“Common” is measured: kept by at least 6 in 10 of the 175 edited versions of this file we could read. “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

Humanizer

Edit the prose, not the author's identity. Prefer a small faithful rewrite over a conspicuously "human" performance.

Priority

Apply these rules in order:

  1. Follow the user's explicit wording, tone, language, and format requirements.
  2. Match any supplied writing sample, including its quirks and punctuation.
  3. Preserve meaning, facts, uncertainty, and required harness structure.
  4. Apply the heuristics below only where they improve the text.

Protect before rewriting

Do not alter or invent:

  • names, facts, numbers, dates, quotes, citations, URLs, and claim strength;
  • code, commands, paths, identifiers, logs, frontmatter, data, or link targets;
  • Markdown directives, writing-block fences, Git directives, and inline comments;
  • required output-template sections or the final Quick Recap status and meaning.

Keep domain terms when they are precise. Do not replace useful technical language with casual approximations.

Rewrite pass

  1. Identify the audience, register, language, and intended outcome.
  2. Look for clusters of problems, not isolated words or punctuation:
    • filler, throat-clearing, signposting, and repeated conclusions;
    • inflated significance, sales language, or generic positivity;
    • vague authorities, unsupported certainty, or fabricated specificity;
    • uniform sentence length, forced groups of three, synonym cycling, or stacked punchline fragments;
    • abstract nouns and elaborate verbs where a concrete subject and simple verb are clearer;
    • diff narration in timeless documentation;
    • chatbot residue such as praise, offers to continue, or ceremonial framing.
  3. Rewrite only the affected passages. Keep strong, distinctive phrasing.
  4. Vary rhythm naturally, but do not manufacture slang, jokes, opinions, hesitations, or personal anecdotes.
  5. Read the result once for voice and once for factual fidelity.

An isolated em dash, transition word, polished sentence, or formal term is not evidence of AI writing. Preserve it when it fits the author's voice.

Harness modes

Pasted text

Return only the final rewrite unless the user asks for an audit, alternatives, or an explanation.

File

Edit prose in place. Leave protected content untouched. Report the changed file and the verification performed instead of pasting the whole file.

Embedded

When another task invokes this skill for an email, document, PR description, commit message, or final response, apply the pass silently and return only the requested artifact. Preserve the owning workflow's output template.

Compatibility

  • Run factual, security, and reasoning checks before this style pass.
  • Do not soften criticism, hide uncertainty, or turn evidence into opinion.
  • Keep concise harness responses concise. Natural does not mean longer.
  • Do not add a "humanization audit" unless requested.

Final check

Confirm that:

  • every factual claim still matches the source;
  • no protected token or required section changed accidentally;
  • the result remains in the source language unless translation was requested;
  • the voice fits the author and audience;
  • the rewrite removed actual patterns rather than merely swapping synonyms.

Provenance

Adapted and substantially rewritten for MythOs Aegis from blader/humanizer. The retained MIT notice is in THIRD_PARTY_NOTICES.md.

Shared output profile

Compose the artifact profile from output-templates. Return the rewrite as the deliverable while preserving facts, protected structure, and voice. Preserve stricter domain rules above, omit empty components, and never fill a component with invented evidence.

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.