The Humanizer Skill

$9.99 $49.99
PromptLeadz · Skill Edition

The Humanizer Skill

Strip AI tells from any draft. Match any voice. The skill your readers cannot detect.

Single SKILL.md file Claude · ChatGPT · Gemini · Grok Lifetime updates

Most AI writing screams "AI wrote this". This skill removes the scream, in one pass.

Before

We are excited to announce a revolutionary new feature that will fundamentally transform how our customers navigate the complexities of modern workflow management. This game changing capability leverages cutting edge AI to deliver unprecedented value across the entire organization.

After

We shipped a new feature today. It changes how teams manage their workflow. It uses AI to do the boring parts for you.

Before

Great question! I'd be happy to help with that. It's worth noting that there are several important factors to consider here. Firstly, you'll want to ensure that you're carefully evaluating your options. Secondly, it's crucial to remember that timing plays a key role. Hopefully this is helpful! Let me know if you need anything else.

After

Two things matter here. Pick the option that fits the constraint you cannot change. Move when the cost of waiting is higher than the cost of being wrong.

Before

In today's fast paced world, it is essential to delve into the topic of remote work. Studies show that remote workers are significantly more productive than their office based counterparts. Moreover, companies that embrace remote work tend to attract top tier talent. In conclusion, this shows the importance of rethinking how we work.

After

Remote work is not a perk anymore. It is the default for the best talent. Companies that still demand office attendance are pricing themselves out of the hiring market they want.

SMode 01

Surface tells

Cuts banned vocabulary and stock phrases on sight. Delve. Leverage. Robust. Seamless. The entire AI dictionary, deleted.

TMode 02

Tone

Kills sycophancy at the open. Kills the offer to keep chatting at the close. Drops corporate register one notch.

RMode 03

Rhythm

Varies sentence length. Breaks the rule of three reflex. Inserts the short sentence after the long one.

IMode 04

Imitate voice

Mirrors a writing sample on three axes: sentence length, register, rhythm. Never copies surface tics.

PMode 05

Pare back

Cuts padding, hedges, self reference, knowledge cutoff disclaimers, and the very, really, quite, somewhat infestation.

01Banned vocabulary (delve, leverage, robust, seamless)
02Sycophantic opener (Great question, Happy to help)
03Generic closer (Let me know, Feel free to ask)
04Hedging filler (It is worth noting, It bears mentioning)
05Em dash reflex (decorative pauses on every line)
06Rule of three reflex (every list of exactly three items)
07Inflated significance (crucial, essential, paramount)
08Promotional cliche (revolutionary, game changing)
09Vague attribution (experts say, studies show)
10Tailing negation (this is great, but not perfect)
11Negative parallelism (not just X but Y, repeated)
12Signposting (firstly, secondly, furthermore)
13Fragmented headers (a header every two sentences)
14Ing nominalization (the using of, the implementing of)
15Passive voice overuse (over 25 percent of sentences)
16Subjectless fragment (sentences with no clear subject)
17Generic positive conclusion (In conclusion, this shows)
18Knowledge cutoff disclaimer (As of my last update)
19Repetitive structure (every paragraph same length)
20Transitional crutch (however, moreover, additionally)
21Corporate hedging (could potentially, may possibly)
22Bullet padding (prose mechanically broken into bullets)
23Stem repetition (three sentences starting the same way)
24Over qualification (very, really, quite, basically)
25Cliched metaphor (north star, table stakes)
26Authority trope (as Aristotle said, as the saying goes)
27False precision (exactly 73 percent, no source)
28Self reference (as an AI, as a language model)
29The however reflex (every counterpoint opens however)
Claude Code Claude Desktop ChatGPT Gemini Grok Cursor Codex CLI Windsurf Cline OpenCode

SKILL.md is the open standard for AI coding assistants. For consumer chat tools (ChatGPT, Gemini, Grok, Claude), paste the file as a system prompt or custom instruction. The install takes under a minute.

01

The STRIP Framework

Five mode router with rules for each. Run the modes in order, S and T first, R, I, P after.

02

The Banned Lists

Explicit word and phrase ban lists for vocabulary, openers, closers, hedges, and connectives. Concrete, not vague.

03

The 29 Tells Checklist

Numbered detection rules. Every AI pattern named, every fix specified. Use as a self audit, not just a one shot pass.

04

Voice Matching Protocol

How to mirror a writing sample on sentence length, register, and rhythm without copying surface tics.

05

When Not To Humanize

Legal, regulatory, academic, and translation contexts where AI tells are actually correct. Explicit handoff rules.

06

Calibration Modes

Aggressive and subtle dials. Plus the ship test: pass the output back through and changes should be minimal.

Will my writing pass AI detectors?

AI detectors are unreliable across the board, so the right test is whether a careful human reader can tell. After one pass with this skill, most readers cannot. After two passes with a voice sample provided, almost none can. The skill optimises for human readability, not for fooling detectors.

How is this different from asking the AI to "make it sound human"?

That prompt usually swaps one set of AI tells for another (more cliches, more sycophancy, longer sentences). This skill bans 29 specific patterns by name, with concrete word and phrase lists. It is detection plus replacement, not a vague vibe.

Does it work outside Claude?

Yes. SKILL.md is the open standard for AI coding assistants: Claude Code, Codex CLI, Cursor, Windsurf, Cline, OpenCode. For consumer chat tools, paste the file as a system prompt or custom instruction. ChatGPT, Gemini, Grok, and Claude all accept the format.

Can I match my own voice?

Yes. Paste a writing sample (one or two paragraphs is enough) alongside the draft and the skill mirrors sentence length, register, and rhythm. It does not copy your surface tics. The output reads as you, not as a parody of you.

Pairs well with The Pilot Skill?

Yes. Pilot tells the AI what to write. Humanizer makes the output sound like you wrote it. Use them together: Pilot drafts, Humanizer polishes. The two skills do not overlap; they compose.

What format do I download?

A single SKILL.md file. Plain markdown. Open in any text editor. Drop in your skills directory or paste as a system prompt. Done.

Lifetime updates?

Yes. One purchase. Free updates as new AI tells emerge and the detection list evolves.

· · ·

Most AI writing reads like AI writing. After one pass, this does not.

PROMPTLEADZ · THE HUMANIZER SKILL · STRIP THE TELLS

Bargain amount
You've got 3 shots to bargain, so use them wisely!