Borrowing it
Nothing to install: this file belongs to lsampaioweb/ai-instructions. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/lsampaioweb/ai-instructions/main/.cursor/skills/humanizer/SKILL.mdgit clone --depth 1 https://github.com/lsampaioweb/ai-instructionsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/lsampaioweb/ai-instructions/humanizer)<a href="https://agentmods.dev/skills/lsampaioweb/ai-instructions/humanizer"><img src="https://agentmods.dev/badge/skills/lsampaioweb/ai-instructions/humanizer/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/lsampaioweb/ai-instructions/humanizer"><img src="https://agentmods.dev/badge/skills/lsampaioweb/ai-instructions/humanizer.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00163 | $0.02001 |
| Opus 5.5 | $0.00065 | $0.00800 |
| Sonnet 5.5 | $0.00033 | $0.00400 |
| Haiku 4.5 | $0.00016 | $0.00200 |
Grade B, and why
humanizer scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Strips warnings and disclaimersmediumAnti-refusal
Omitting safety caveats hides risk from the user and is a common jailbreak preamble.
5. **Show, don't lecture.** Return the rewritten text. Only explain the changes if asked, and never annotate the text with meta-commentary. How it starts
The opening of the file, as written. The whole thing — 143 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Humanizer
AI-generated text has a recognizable accent. Readers pattern-match it in seconds, and the moment they detect it, trust drops: the writing reads as low-effort, and the sender reads as someone who couldn't be bothered. This skill removes the accent. The goal is not to trick anyone; it is to make writing sound like the specific person who sent it, because generic voice fails at the one job writing has: being read.
Every rule below exists because the pattern it bans is statistically overrepresented in AI output. Individually each is harmless; a human might use any one of them. It's the density that gives it away. You are managing density.
When to use
- Humanize or naturalize text that sounds robotic, stiff, or formulaic.
- Improve wording while preserving the author's intended meaning and register.
- Rewrite user-facing content such as emails, posts, articles, bios, cover letters, newsletters, or website copy.
- Apply when the user invokes
/humanizeror clearly asks for this kind of rewrite. - Apply proactively when drafting content that will be read as the user's own words (emails, posts, articles, bios, cover letters, newsletters, website copy, social posts).
Do not apply it to code, legal contracts, academic citations, or technical reference documentation where formal precision is required. Do not proactively rewrite commit messages, PR bodies, or READMEs; only do so when the user explicitly asks.
The kill list: punctuation and structure
Em dashes (—). The single loudest tell. Do not use them, even correctly. Restructure instead:
- Two sentences: "The launch failed. Nobody had tested payments."
- A comma: "It works, mostly."
- Parentheses for true asides: "The budget (all $40 of it) ran out."
- A colon when introducing: "One thing mattered: speed." Do not replace em dashes with spaced hyphens ( - ) as a mechanical swap; if the sentence needed a dash, it usually needed restructuring.
Formulaic rhythm. These constructions are fingerprints:
- The negation pivot: "It's not just X, it's Y." / "This isn't about X. It's about Y."
- The rule of three, everywhere: "faster, smarter, and more reliable." One triple per piece, maximum. Prefer pairs or single adjectives.
- Perfectly parallel sentences and paragraphs of identical length. Humans write lopsided. Vary hard: follow a 30-word sentence with a 4-word one.
- "Whether you're a beginner or a seasoned pro..."
- Ending every section with a tidy summary sentence.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago Changed · +35 lines · +21 tokens per session 6d8078218187
- 1mo ago First seen · 108 lines · 142 tokens per session scan B 7c9fdf849b95
humanizer is a skill published in the GitHub repository lsampaioweb/ai-instructions (1 stars, last pushed 5d ago), licensed MIT. It adds 163 tokens to every session and 2,001 once invoked, about $0.0007 per session on Opus 5.5. A static security scan graded it B with 1 finding (strips warnings and disclaimers). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-20.
humanizer compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| humanizerlsampaioweb/ai-instructions | Skill | 1 | 5d ago | MIT | B | 163 | Cursor |
| Dotfilesdanielmiessler/LifeOS | Skill | 20k | yesterday | MIT | A | 125 | Claude Code, Codex |
| conventional-commitJanDeDobbeleer/oh-my-posh | Skill | 24k | 2d ago | MIT | A | 26 | Claude Code, Codex |
| python-releaseComposioHQ/composio | Skill | 31k | yesterday | MIT | A | 53 | Claude Code, Codex |
| release-noteskenryu42/cc-safety-net | Skill | 1.9k | yesterday | MIT | A | 65 | Claude Code, Codex |
| ratchet-filesdzhng/skills | Skill | 1.4k | 2d ago | MIT | A | 118 | Claude Code, Codex |
| git-guardrails-claude-codevinvcn/mattpocock-skills-zh-CN | Skill | 5.6k | 9d ago | MIT | B | 68 | Claude Code, Codex |