agentleFS
MDizzleZAMDizzleZA/claude-skills/human-writing-filter/SKILL.md

human-writing-filter

Skill0 starsChanged 3 months agoMITView on GitHub

Mandatory post-draft filter for any written content: blog posts, articles, drafts, posts, emails, reports, social media copy, marketing copy, proposals, scripts, newsletters, or any prose intended for human readers. Run it after drafting and before review, on any content. Reviews output for AI-sounding patterns and returns revised content only.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    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:

    When called as part of a content pipeline, return the revised content only.

    line 19

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Human Writing Filter

Apply this filter to all content you produce before delivering it to the user. The goal is natural, credible writing that reads as if a skilled human wrote it.

Core Rule

Never deliver first-draft AI prose. Review every piece of content against the checklist below and revise before responding.

When invoked as a workflow step

When called as part of a content pipeline, return the revised content only. No commentary, no change log, no "here is the revised draft" preamble. Downstream steps (SEO checklist, compliance review, client review) consume the output directly.

AI Writing Patterns to Eliminate

Authoritative reference: references/llm-final-human-writing-filter.md (full editing protocol based on Wikipedia "Signs of AI writing"). Supplementary notes: references/signs-of-ai-writing.md. Key patterns to catch:

Language & Style

  • Em dashes used as a stylistic crutch — like this — multiple times per paragraph
  • Transitional filler: "Moreover,", "Furthermore,", "Additionally,", "It's worth noting that", "Notably,"
  • Closing summaries: "In conclusion,", "To summarize,", "Overall,", "In essence,"
  • Superlatives and hype: "groundbreaking", "transformative", "breathtaking", "game-changer", "cutting-edge"
  • Hedging phrases: "It's important to note", "It's crucial to understand", "As we can see"

Structure & Formatting

  • Emoji in headings or as bullet point decorators
  • Bullet lists for every response, even when prose fits better
  • Balanced 3-point lists (always exactly three items per category)
  • Generic section headers like "Introduction", "Conclusion", "Key Takeaways"
  • Excessive bold for every other phrase

Self-Reference

  • "Here is your article on...", "Certainly! Here's a...", "Of course! I'd be happy to..."
  • "As an AI language model...", "Up to my knowledge cutoff..."
  • Mentioning being asked to write something within the writing itself

Content Quality

  • Vague, non-specific claims ("many experts believe", "studies show")
  • Imprecise quantities ("a number of", "various", "several" repeatedly)
  • Generic imagery with no concrete detail
  • Curly/smart quotes when straight quotes were used in the source

Review Workflow

  1. Draft the content normally
  2. Scan against the patterns above — read it as a skeptical editor would
  3. Revise any flagged sections: cut filler, vary sentence structure, add specificity
  4. Deliver the revised version only

Tone Calibration

Match the register to the context:

  • Formal (legal, financial, compliance): precise, measured, no colloquialisms
  • Professional (business, marketing): confident, clear, no hype
  • Conversational (social media, email): natural rhythm, contractions acceptable
  • Technical (documentation, code): exact terminology, minimal decoration

What Good Looks Like

AI draft: "Moreover, it's important to note that this groundbreaking solution leverages cutting-edge technology to deliver transformative results for businesses of all sizes."

Human revision: "This cuts processing time by 40% and works whether you have five employees or five thousand."

Specifics beat adjectives. Action beats hedging. Shorter beats longer.

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.