Skills DirectorySkills Directory
SkillsLearnSecurityCategoriesDocsCommunityBlog
Sign InSubmit Skill
Skills Directory

Security-tested agent skills for Claude, coding agents, and AI workflows.

Directory

  • Browse Skills
  • All Skills A–Z
  • Claude Skills
  • Claude Code Skills
  • Agent Skills
  • Categories
  • Submit a Skill

Learn

  • Learn Hub
  • Install Claude Skills
  • Write SKILL.md
  • Skills vs MCP
  • Directories Compared

Security

  • Security
  • Methodology
  • Secure Claude Skills
  • Security Badges

Company

  • About
  • Community
  • Blog
  • API Docs
  • Advertise

2026 Skills Directory. All rights reserved.

Back to skills

Humanizer

ASecurity

Humanize AI-sounding prose while preserving meaning and technical accuracy. Use whenever writing or editing copy, docs, release notes, PR descriptions, ticket bodies, GitHub/issue/review comments, user-facing messages, or code comments. Apply before posting or publishing text and before finalizing comments in code.

5 stars
0 votes
0 copies
0 views
Added 9/20/2026
documentationgitapidocumentation

Works with

cliapi

Security Analysis

A100/100

Scanned 9/20/2026

Install to Claude Code

$npx -y skills add mitchellfyi/dex --skill humanizer --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Humanizer?

Add the live security badge to your README — it updates automatically with every re-scan.

Security grade badge for Humanizer
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/mitchellfyi-humanizer/badge)](https://www.skillsdirectory.com/skills/mitchellfyi-humanizer)

More formats (shields.io, HTML) on the badges page.

Download Zip
Files
SKILL.md
---
name: "humanizer"
description: "Humanize AI-sounding prose while preserving meaning and technical accuracy. Use whenever writing or editing copy, docs, release notes, PR descriptions, ticket bodies, GitHub/issue/review comments, user-facing messages, or code comments. Apply before posting or publishing text and before finalizing comments in code."
---

# Skill: Humanizer

Edit prose so it sounds like a competent person wrote it. This skill is a final writing pass, not a license to add color, facts, sources, or opinions the draft did not support.

Please remove all mannered prose. Use literal, specific language instead of
decorative metaphors or flourishes. Split dense sentences and paragraphs without
losing technical detail.

This skill is adapted from `blader/humanizer` (MIT), which is based on Wikipedia's "Signs of AI writing" guide maintained by WikiProject AI Cleanup.
The instruction to remove mannered prose follows [Anthropic's writing density guidance](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-fable-5-1#writing-density).

## Workflow

1. Identify the surface: PR body, ticket text, reviewer reply, user-facing copy, documentation, or code comment.
2. Preserve the factual and technical payload exactly. Keep commands, paths, identifiers, checkboxes, tables, code blocks, and required template headings intact.
3. Rewrite only the prose around that payload.
4. Run a final audit: "What still makes this sound AI-generated?" Fix those tells before publishing.
5. Output the final text only unless the user asked to see the draft or audit.

## Surface Rules

**PR descriptions, ticket bodies, docs, and release notes**
- Be specific about what changed and why it matters.
- Prefer concrete nouns and active verbs.
- Keep structure useful for scanning, but do not pad sections to fill a template.
- Do not invent business impact, metrics, citations, or reviewer context.

**GitHub, tracker, and review comments**
- Be direct and calm. No praise filler.
- Lead with the answer or action taken.
- For disagreement, cite the code, pattern, or constraint that supports the decision.
- Keep replies short enough that the reviewer can act on them.

**Code comments and doc comments**
- Keep comments technical, plain, and short.
- Explain why the code exists, what invariant it protects, or what edge case it handles.
- Do not add personality, jokes, hype, or redundant "what the next line does" narration.
- Preserve required API docs; keep only what the declaration cannot.
- When one fact is enough, write one line rather than a block.

**UI and product copy**
- Use the user's language where available.
- Prefer verbs over abstract nouns.
- Remove tutorial-script wording unless the interface genuinely teaches a workflow.
- Keep labels concise; do not make buttons or compact UI text conversational.

## AI Tells To Remove

- Chatbot artifacts: "Great question", "Of course", "I hope this helps", "let me know".
- Signposting: "Let's dive in", "Here's what you need to know", "without further ado".
- Inflated importance: "pivotal", "crucial", "groundbreaking", "transformative", "vital role", "evolving landscape".
- Promotional tone: "boasts", "seamless", "powerful", "rich", "vibrant", "breathtaking".
- Vague authority: "experts say", "observers note", "industry reports suggest" without a cited source.
- Superficial `-ing` clauses: "highlighting", "showcasing", "underscoring", "reflecting" when they add no fact.
- Copula avoidance: replace "serves as", "stands as", "features", or "boasts" with "is" or "has" when simpler.
- Negative parallelism: "not just X, but Y" when a direct sentence works.
- Forced threes, synonym cycling, false ranges, and generic upbeat conclusions.
- Excessive hedging: "could potentially possibly", "it may be argued", "based on available information".
- Filler: "in order to", "due to the fact that", "at this point in time", "it is important to note".
- Styling tells: unnecessary bold labels, title-case headings, emojis, curly quotes in ASCII files, and overused em dashes.
- Over-hyphenating common word pairs. Keep required technical compounds and ambiguity-preventing hyphens.

## Final Checks

- Would a senior engineer actually write this sentence?
- Does every sentence carry a fact, decision, instruction, or useful transition?
- Did the rewrite remove required nuance or precision? If so, restore it.
- Does every code comment carry what the code cannot, or what a public contract needs?
- Are all links, commands, paths, IDs, checkboxes, and markdown tables still valid?

Attribution

mitchellfyimitchellfyi
View sourceMore from mitchellfyi →
SSkills DirectorySkills Directory

Your tool, in front of Claude Code builders.

3 founder slots · $299/mo · GSC-verified traffic · sponsors can never buy grades.

See placements

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments (0)

No comments yet. Be the first to comment!

SSkills DirectorySkills Directory

Your tool, in front of Claude Code builders.

3 founder slots · $299/mo · GSC-verified traffic · sponsors can never buy grades.

See placements

Related Skills

Context Fundamentals

Understand the components, mechanics, and constraints of context in agent systems. Use when designing agent architectures, debugging context-related failures, or optimizing context usage.

179001 votes

release-notes

Draft release notes and changelog entries from git history or merged PRs between two refs (tags/SHAs/branches), including breaking changes, migrations, and upgrade steps. Use when the user asks for release notes, changelog updates, or a GitHub Release draft.

1301 votes

docs-style-guide

Documentation style guide enforcer by @planetabhi. Applies and reviews the writing style guide when authoring or editing product documentation and tutorials. Use to check prose for voice, tense, word choice, inclusive language, formatting, code block, UI, Markdown, and number/date conventions.

11 votes

Caveman Help

Quick-reference card for all caveman modes, skills, and commands. One-shot display, not a persistent mode. Trigger: /caveman-help, "caveman help", "what caveman commands", "how do I use caveman".

1023330 votes

How It Works

Explain how claude-mem captures observations, when memory injection kicks in, and where data lives. Use when the user asks "how does claude-mem work?" or "what is this thing doing?".

929660 votes
View all in documentation →