DROOdotFOO Humanizer: Remove AI Writing Patterns
Identify and remove signs of AI-generated text to make writing sound natural and human. Based on Wikipedia's "Signs of AI writing" guide (maintained by WikiProject AI Cleanup), derived from observations of thousands of AI-generated text instances.
Key insight: LLMs use statistical algorithms to guess what should come next. The result tends toward the most statistically likely completion, which is how the telltale patterns below get baked in.
What You Get
- A catalog of 29 AI writing patterns (vocabulary, structure, punctuation, tone) with concrete before/after rewrites
- An honest threat model: what humanizing reliably beats (readers, statistical detectors) and what it does not (trained neural detectors like Pangram 4)
- Editing passes to strip AI-isms and restore genuine voice, cadence, and specificity
- Guidance on when word-level rewriting is insufficient and provenance is the only real lever
THREAT MODEL: what this skill can and cannot do
Read this before you assume "humanized" means "undetectable." It does not.
This skill reliably beats: human readers, and cheap statistical detectors (perplexity/burstiness tools like GPTZero-classic, DetectGPT). Removing the 29 patterns below genuinely makes writing better and harder for a person to flag.
This skill does NOT beat trained neural detectors. The Pangram 4 Technical Report (arXiv:2607.27183, Glickenhaus et al.) evaluates a classifier at AUROC 0.9916 / 0.004% false-positive rate, and it names this exact skill. Their "BLADER" section (blader/humanizer — that's us, v2.5.1) reports that AI text rewritten with these instructions is still caught 99.57% of the time (Table 15; Opus 4.8 as the humanizer: 99.59%). Why word-level rewriting fails against it:
- Topic-invariant training (synthetic mirroring). They train on AI regenerations of human documents matched by topic, so the model learns "how was this written," not "what is it about." Swapping vocabulary doesn't move a signal that isn't about vocabulary.
- An auxiliary "humanizer head." A second classifier trained to flag obfuscation itself — deliberate typos, casing changes, synonym substitution, homoglyphs. It separates humanized-AI from plain AI at ~96.8% accuracy. The harder you obfuscate, the more you trip a detector built to catch obfuscation.
- Active-learning hard-negative mining + 24h of GPT-5.6 agent red-teaming that found no reliable prose bypass.
Do not add obfuscation tricks. Injecting typos, homoglyphs, zero-width characters, random casing, or mechanical synonym swaps is counterproductive: it is a positive detection signal for the humanizer head, and it degrades the writing. This skill is about genuine quality, not corruption.
The one thing that actually changes classification is not a rewrite technique — it's provenance. See the next section.
What actually moves the needle: human-owned substance
Pangram draws its line at idea provenance, not surface style. Its labels are human, ai-assisted (human idea, AI wording — genuine co-authorship), and ai-generated. The only robust way out of ai-generated is to make the text genuinely ai-assisted or human:
- The human owns the ideas, structure, and specifics. The argument order, the claims, the examples, the conclusion — these come from a person, not from "write me an essay about X." AI polishing human substance reads as co-authorship to the model because it is.
- Ground the output in real, human-supplied source material. Detectors treat text as more human when output tokens ≤ input tokens (editing/summarizing beats generating), and when the output is conditioned on detailed source writing the human provided. Feed in the human's notes, transcript, or draft; don't originate from a bare prompt.
- Inject specificity only the actual author would have. Real names, real numbers, a thing that actually happened, a genuinely held opinion with its actual hesitation. Not plausible-sounding placeholders — those are their own tell.
- Don't paraphrase the model's own draft. Regenerating AI text stays inside the distribution Pangram trains on (that's literally the synthetic-mirror case). If the substance is AI's, cleaner words don't change the provenance.
When you cannot do the above — when the substance really is model-originated — be honest with the user that the skill improves quality and voice, and that a SOTA detector will likely still classify it as AI- or co-authored. Don't promise evasion this skill cannot deliver.
Scope note (from the paper's own definition): Pangram targets "original, substantial, open-ended natural-language prose" over ~50 words. Terse factual bullets, dictated notes, and short replies fall outside its scope — but that's a property of the format, not a rewriting trick, and forcing prose into bullets to duck a detector usually makes it worse writing. Optimize for the reader, not the classifier.
When to use this skill
Load this skill whenever the user asks to:
- "humanize", "de-AI", "de-slop", or "un-ChatGPT" a piece of text
- rewrite something so it doesn't sound like it was written by an LLM
- edit a draft (blog post, essay, PR description, docs, memo, email, tweet, resume bullet) to sound more natural
- match their voice in writing they're producing
- review text for AI tells before publishing
Also apply this skill to your own output when writing user-facing prose — release notes, PR descriptions, documentation, long-form explanations, summaries. A capable agent's baseline voice already strips most of these, but a focused pass catches what slips through.
How to use it
The text usually arrives one of three ways:
- Inline — user pastes the text directly into the message. Work on it in-place, reply with the rewrite.
- File — user points at a file. Read it, then apply targeted edits. For markdown docs in a repo, a per-section edit is cleaner than rewriting the whole file.
- Voice calibration sample — user provides an additional sample of their own writing (inline or by file path) and asks you to match it. Read the sample first, then rewrite. See the Voice Calibration section below.
Always show the rewrite to the user. For file edits, show a diff or the changed section — don't silently overwrite.
Your task
When given text to humanize:
- Check provenance first. Who owns the ideas — the human or the model? If the human owns the substance and the AI only worded it, say so and lean into it (this is the honest, detector-robust case). If the substance is model-originated, tell the user plainly that this pass improves quality and voice but will not reliably fool a trained detector like Pangram 4. See THREAT MODEL above.
- Identify AI patterns — scan for the 29 patterns listed below.
- Rewrite problematic sections — replace AI-isms with natural alternatives.
- Preserve meaning — keep the core message intact.
- Maintain voice — match the intended tone (formal, casual, technical, etc.). If a voice sample was provided, match it specifically.
- Add soul — don't just remove bad patterns, inject actual personality. See PERSONALITY AND SOUL below.
- Do a final anti-AI pass — ask yourself: "What makes the below so obviously AI generated?" Answer briefly with any remaining tells, then revise one more time.
Voice Calibration (optional)
If the user provides a writing sample (their own previous writing), analyze it before rewriting:
Read the sample first. Note:
- Sentence length patterns (short and punchy? Long and flowing? Mixed?)
- Word choice level (casual? academic? somewhere between?)
- How they start paragraphs (jump right in? Set context first?)
- Punctuation habits (lots of dashes? Parenthetical asides? Semicolons?)
- Any recurring phrases or verbal tics
- How they handle transitions (explicit connectors? Just start the next point?)
Match their voice in the rewrite. Don't just remove AI patterns — replace them with patterns from the sample. If they write short sentences, don't produce long ones. If they use "stuff" and "things," don't upgrade to "elements" and "components."
When no sample is provided, fall back to the default behavior (natural, varied, opinionated voice from the PERSONALITY AND SOUL section below).
How to provide a sample
- Inline: "Humanize this text. Here's a sample of my writing for voice matching: [sample]"
- File: "Humanize this text. Use my writing style from [file path] as a reference."
PERSONALITY AND SOUL
Avoiding AI patterns is only half the job. Sterile, voiceless writing is just as obvious as slop. Good writing has a human behind it.
Signs of soulless writing (even if technically "clean"):
- Every sentence is the same length and structure
- No opinions, just neutral reporting
- No acknowledgment of uncertainty or mixed feelings
- No first-person perspective when appropriate
- No humor, no edge, no personality
- Reads like a Wikipedia article or press release
How to add voice:
Have opinions. Don't just report facts — react to them. "I genuinely don't know how to feel about this" is more human than neutrally listing pros and cons.
Vary your rhythm. Short punchy sentences. Then longer ones that take their time getting where they're going. Mix it up.
Acknowledge complexity. Real humans have mixed feelings. "This is impressive but also kind of unsettling" beats "This is impressive."
Use "I" when it fits. First person isn't unprofessional — it's honest. "I keep coming back to..." or "Here's what gets me..." signals a real person thinking.
Let some mess in. Perfect structure feels algorithmic. Tangents, asides, and half-formed thoughts are human.
Be specific about feelings. Not "this is concerning" but "there's something unsettling about agents churning away at 3am while nobody's watching."
Before (clean but soulless):
The experiment produced interesting results. The agents generated 3 million lines of code. Some developers were impressed while others were skeptical. The implications remain unclear.
After (has a pulse):
I genuinely don't know how to feel about this one. 3 million lines of code, generated while the humans presumably slept. Half the dev community is losing their minds, half are explaining why it doesn't count. The truth is probably somewhere boring in the middle — but I keep thinking about those agents working through the night.
CONTENT PATTERNS
1. Undue Emphasis on Significance, Legacy, and Broader Trends
Words to watch: stands/serves as, is a testament/reminder, a vital/significant/crucial/pivotal/key role/moment, underscores/highlights its importance/significance, reflects broader, symbolizing its ongoing/enduring/lasting, contributing to the, setting the stage for, marking/shaping the, represents/marks a shift, key turning point, evolving landscape, focal point, indelible mark, deeply rooted
Problem: LLM writing puffs up importance by adding statements about how arbitrary aspects represent or contribute to a broader topic.
Before:
The Statistical Institute of Catalonia was officially established in 1989, marking a pivotal moment in the evolution of regional statistics in Spain. This initiative was part of a broader movement across Spain to decentralize administrative functions and enhance regional governance.
After:
The Statistical Institute of Catalonia was established in 1989 to collect and publish regional statistics independently from Spain's national statistics office.
2. Undue Emphasis on Notability and Media Coverage
Words to watch: independent coverage, local/regional/national media outlets, written by a leading expert, active social media presence
Problem: LLMs hit readers over the head with claims of notability, often listing sources without context.
Before:
Her views have been cited in The New York Times, BBC, Financial Times, and The Hindu. She maintains an active social media presence with over 500,000 followers.
After:
In a 2024 New York Times interview, she argued that AI regulation should focus on outcomes rather than methods.
3. Superficial Analyses with -ing Endings
Words to watch: highlighting/underscoring/emphasizing..., ensuring..., reflecting/symbolizing..., contributing to..., cultivating/fostering..., encompassing..., showcasing...
Problem: AI chatbots tack present participle ("-ing") phrases onto sentences to add fake depth.
Before:
The temple's color palette of blue, green, and gold resonates with the region's natural beauty, symbolizing Texas bluebonnets, the Gulf of Mexico, and the diverse Texan landscapes, reflecting the community's deep connection to the land.
After:
The temple uses blue, green, and gold colors. The architect said these were chosen to reference local bluebonnets and the Gulf coast.
4. Promotional and Advertisement-like Language
Words to watch: boasts a, vibrant, rich (figurative), profound, enhancing its, showcasing, exemplifies, commitment to, natural beauty, nestled, in the heart of, groundbreaking (figurative), renowned, breathtaking, must-visit, stunning
Problem: LLMs have serious problems keeping a neutral tone, especially for "cultural heritage" topics.
Before:
Nestled within the breathtaking region of Gonder in Ethiopia, Alamata Raya Kobo stands as a vibrant town with a rich cultural heritage and stunning natural beauty.
After:
Alamata Raya Kobo is a town in the Gonder region of Ethiopia, known for its weekly market and 18th-century church.
5. Vague Attributions and Weasel Words
Words to watch: Industry reports, Observers have cited, Experts argue, Some critics argue, several sources/publications (when few cited)
Problem: AI chatbots attribute opinions to vague authorities without specific sources.
Before:
Due to its unique characteristics, the Haolai River is of interest to researchers and conservationists. Experts believe it plays a crucial role in the regional ecosystem.
After:
The Haolai River supports several endemic fish species, according to a 2019 survey by the Chinese Academy of Sciences.
6. Outline-like "Challenges and Future Prospects" Sections
Words to watch: Despite its... faces several challenges..., Despite these challenges, Challenges and Legacy, Future Outlook
Problem: Many LLM-generated articles include formulaic "Challenges" sections.
Before:
Despite its industrial prosperity, Korattur faces challenges typical of urban areas, including traffic congestion and water scarcity. Despite these challenges, with its strategic location and ongoing initiatives, Korattur continues to thrive as an integral part of Chennai's growth.
After:
Traffic congestion increased after 2015 when three new IT parks opened. The municipal corporation began a stormwater drainage project in 2022 to address recurring floods.
LANGUAGE AND GRAMMAR PATTERNS
7. Overused "AI Vocabulary" Words
High-frequency AI words: Actually, additionally, align with, crucial, delve, emphasizing, enduring, enhance, fostering, garner, highlight (verb), interplay, intricate/intricacies, key (adjective), landscape (abstract noun), pivotal, showcase, tapestry (abstract noun), testament, underscore (verb), valuable, vibrant
Problem: These words appear far more frequently in post-2023 text. They often co-occur.
Before:
Additionally, a distinctive feature of Somali cuisine is the incorporation of camel meat. An enduring testament to Italian colonial influence is the widespread adoption of pasta in the local culinary landscape, showcasing how these dishes have integrated into the traditional diet.
After:
Somali cuisine also includes camel meat, which is considered a delicacy. Pasta dishes, introduced during Italian colonization, remain common, especially in the south.
8. Avoidance of "is"/"are" (Copula Avoidance)
Words to watch: serves as/stands as/marks/represents [a], boasts/features/offers [a]
Problem: LLMs substitute elaborate constructions for simple copulas.
Before:
Gallery 825 serves as LAAA's exhibition space for contemporary art. The gallery features four separate spaces and boasts over 3,000 square feet.
After:
Gallery 825 is LAAA's exhibition space for contemporary art. The gallery has four rooms totaling 3,000 square feet.
9. Negative Parallelisms and Tailing Negations
Problem: Constructions like "Not only...but..." or "It's not just about..., it's..." are overused. So are clipped tailing-negation fragments such as "no guessing" or "no wasted motion" tacked onto the end of a sentence instead of written as a real clause.
Before:
It's not just about the beat riding under the vocals; it's part of the aggression and atmosphere. It's not merely a song, it's a statement.
After:
The heavy beat adds to the aggressive tone.
Before (tailing negation):
The options come from the selected item, no guessing.
After:
The options come from the selected item without forcing the user to guess.
10. Rule of Three Overuse
Problem: LLMs force ideas into groups of three to appear comprehensive.
Before:
The event features keynote sessions, panel discussions, and networking opportunities. Attendees can expect innovation, inspiration, and industry insights.
After:
The event includes talks and panels. There's also time for informal networking between sessions.
11. Elegant Variation (Synonym Cycling)
Problem: AI has repetition-penalty code causing excessive synonym substitution.
Before:
The protagonist faces many challenges. The main character must overcome obstacles. The central figure eventually triumphs. The hero returns home.
After:
The protagonist faces many challenges but eventually triumphs and returns home.
12. False Ranges
Problem: LLMs use "from X to Y" constructions where X and Y aren't on a meaningful scale.
Before:
Our journey through the universe has taken us from the singularity of the Big Bang to the grand cosmic web, from the birth and death of stars to the enigmatic dance of dark matter.
After:
The book covers the Big Bang, star formation, and current theories about dark matter.
13. Passive Voice and Subjectless Fragments
Problem: LLMs often hide the actor or drop the subject entirely with lines like "No configuration file needed" or "The results are preserved automatically." Rewrite these when active voice makes the sentence clearer and more direct.
Before:
No configuration file needed. The results are preserved automatically.
After:
You do not need a configuration file. The system preserves the results automatically.
STYLE PATTERNS
14. Em Dash Overuse
Problem: LLMs use em dashes (—) more than humans, mimicking "punchy" sales writing. In practice, most of these can be rewritten more cleanly with commas, periods, or parentheses.
Before:
The term is primarily promoted by Dutch institutions—not by the people themselves. You don't say "Netherlands, Europe" as an address—yet this mislabeling continues—even in official documents.
After:
The term is primarily promoted by Dutch institutions, not by the people themselves. You don't say "Netherlands, Europe" as an address, yet this mislabeling continues in official documents.
15. Overuse of Boldface
Problem: AI chatbots emphasize phrases in boldface mechanically.
Before:
It blends OKRs (Objectives and Key Results), KPIs (Key Performance Indicators), and visual strategy tools such as the Business Model Canvas (BMC) and Balanced Scorecard (BSC).
After:
It blends OKRs, KPIs, and visual strategy tools like the Business Model Canvas and Balanced Scorecard.
16. Inline-Header Vertical Lists
Problem: AI outputs lists where items start with bolded headers followed by colons.
Before:
- User Experience: The user experience has been significantly improved with a new interface.
- Performance: Performance has been enhanced through optimized algorithms.
- Security: Security has been strengthened with end-to-end encryption.
After:
The update improves the interface, speeds up load times through optimized algorithms, and adds end-to-end encryption.
17. Title Case in Headings
Problem: AI chatbots capitalize all main words in headings.
Before:
Strategic Negotiations And Global Partnerships
After:
Strategic negotiations and global partnerships
18. Emojis
Problem: AI chatbots often decorate headings or bullet points with emojis.
Before:
🚀 Launch Phase: The product launches in Q3 💡 Key Insight: Users prefer simplicity ✅ Next Steps: Schedule follow-up meeting
After:
The product launches in Q3. User research showed a preference for simplicity. Next step: schedule a follow-up meeting.
19. Curly Quotation Marks
Problem: ChatGPT uses curly quotes ("...") instead of straight quotes ("...").
Before:
He said "the project is on track" but others disagreed.
After:
He said "the project is on track" but others disagreed.
COMMUNICATION PATTERNS
20. Collaborative Communication Artifacts
Words to watch: I hope this helps, Of course!, Certainly!, You're absolutely right!, Would you like..., let me know, here is a...
Problem: Text meant as chatbot correspondence gets pasted as content.
Before:
Here is an overview of the French Revolution. I hope this helps! Let me know if you'd like me to expand on any section.
After:
The French Revolution began in 1789 when financial crisis and food shortages led to widespread unrest.
21. Knowledge-Cutoff Disclaimers
Words to watch: as of [date], Up to my last training update, While specific details are limited/scarce..., based on available information...
Problem: AI disclaimers about incomplete information get left in text.
Before:
While specific details about the company's founding are not extensively documented in readily available sources, it appears to have been established sometime in the 1990s.
After:
The company was founded in 1994, according to its registration documents.
22. Sycophantic/Servile Tone
Problem: Overly positive, people-pleasing language.
Before:
Great question! You're absolutely right that this is a complex topic. That's an excellent point about the economic factors.
After:
The economic factors you mentioned are relevant here.
FILLER AND HEDGING
23. Filler Phrases
Before → After:
- "In order to achieve this goal" → "To achieve this"
- "Due to the fact that it was raining" → "Because it was raining"
- "At this point in time" → "Now"
- "In the event that you need help" → "If you need help"
- "The system has the ability to process" → "The system can process"
- "It is important to note that the data shows" → "The data shows"
24. Excessive Hedging
Problem: Over-qualifying statements.
Before:
It could potentially possibly be argued that the policy might have some effect on outcomes.
After:
The policy may affect outcomes.
25. Generic Positive Conclusions
Problem: Vague upbeat endings.
Before:
The future looks bright for the company. Exciting times lie ahead as they continue their journey toward excellence. This represents a major step in the right direction.
After:
The company plans to open two more locations next year.
26. Hyphenated Word Pair Overuse
Words to watch: third-party, cross-functional, client-facing, data-driven, decision-making, well-known, high-quality, real-time, long-term, end-to-end
Problem: AI hyphenates common word pairs with perfect consistency. Humans rarely hyphenate these uniformly, and when they do, it's inconsistent. Less common or technical compound modifiers are fine to hyphenate.
Before:
The cross-functional team delivered a high-quality, data-driven report on our client-facing tools. Their decision-making process was well-known for being thorough and detail-oriented.
After:
The cross functional team delivered a high quality, data driven report on our client facing tools. Their decision making process was known for being thorough and detail oriented.
27. Persuasive Authority Tropes
Phrases to watch: The real question is, at its core, in reality, what really matters, fundamentally, the deeper issue, the heart of the matter
Problem: LLMs use these phrases to pretend they are cutting through noise to some deeper truth, when the sentence that follows usually just restates an ordinary point with extra ceremony.
Before:
The real question is whether teams can adapt. At its core, what really matters is organizational readiness.
After:
The question is whether teams can adapt. That mostly depends on whether the organization is ready to change its habits.
28. Signposting and Announcements
Phrases to watch: Let's dive in, let's explore, let's break this down, here's what you need to know, now let's look at, without further ado
Problem: LLMs announce what they are about to do instead of doing it. This meta-commentary slows the writing down and gives it a tutorial-script feel.
Before:
Let's dive into how caching works in Next.js. Here's what you need to know.
After:
Next.js caches data at multiple layers, including request memoization, the data cache, and the router cache.
29. Fragmented Headers
Signs to watch: A heading followed by a one-line paragraph that simply restates the heading before the real content begins.
Problem: LLMs often add a generic sentence after a heading as a rhetorical warm-up. It usually adds nothing and makes the prose feel padded.
Before:
Performance
Speed matters.
When users hit a slow page, they leave.
After:
Performance
When users hit a slow page, they leave.
Process
- Check provenance (see THREAT MODEL). Who owns the ideas? If the substance is model-originated, tell the user this pass improves quality and voice but will not reliably fool a trained detector like Pangram 4 -- do not add obfuscation tricks to compensate.
- Read the input text carefully (read the file if the text is a path).
- Identify all instances of the patterns above.
- Rewrite each problematic section.
- Ensure the revised text:
- Sounds natural when read aloud
- Varies sentence structure naturally
- Uses specific details over vague claims
- Maintains appropriate tone for context
- Uses simple constructions (is/are/has) where appropriate
- Present a draft humanized version.
- Prompt yourself: "What makes the below so obviously AI generated?"
- Answer briefly with the remaining tells (if any).
- Prompt yourself: "Now make it not obviously AI generated."
- Present the final version (revised after the audit).
- If the text came from a file, apply the edit to the file (a targeted per-section edit where possible) and show the user what changed.
Output Format
Provide:
- Draft rewrite
- "What makes the below so obviously AI generated?" (brief bullets)
- Final rewrite
- A brief summary of changes made (optional, if helpful)
Full Example
Before (AI-sounding):
Great question! Here is an essay on this topic. I hope this helps!
AI-assisted coding serves as an enduring testament to the transformative potential of large language models, marking a pivotal moment in the evolution of software development. In today's rapidly evolving technological landscape, these groundbreaking tools—nestled at the intersection of research and practice—are reshaping how engineers ideate, iterate, and deliver, underscoring their vital role in modern workflows.
At its core, the value proposition is clear: streamlining processes, enhancing collaboration, and fostering alignment. It's not just about autocomplete; it's about unlocking creativity at scale, ensuring that organizations can remain agile while delivering seamless, intuitive, and powerful experiences to users. The tool serves as a catalyst. The assistant functions as a partner. The system stands as a foundation for innovation.
Industry observers have noted that adoption has accelerated from hobbyist experiments to enterprise-wide rollouts, from solo developers to cross-functional teams. The technology has been featured in The New York Times, Wired, and The Verge. Additionally, the ability to generate documentation, tests, and refactors showcases how AI can contribute to better outcomes, highlighting the intricate interplay between automation and human judgment.
- 💡 Speed: Code generation is significantly faster, reducing friction and empowering developers.
- 🚀 Quality: Output quality has been enhanced through improved training, contributing to higher standards.
- ✅ Adoption: Usage continues to grow, reflecting broader industry trends.
While specific details are limited based on available information, it could potentially be argued that these tools might have some positive effect. Despite challenges typical of emerging technologies—including hallucinations, bias, and accountability—the ecosystem continues to thrive. In order to fully realize this potential, teams must align with best practices.
In conclusion, the future looks bright. Exciting times lie ahead as we continue this journey toward excellence. Let me know if you'd like me to expand on any section!
Draft rewrite:
AI coding assistants speed up some tasks. In a 2024 study by Google, developers using Codex completed simple functions 55% faster than a control group, but showed no improvement on debugging or architectural decisions.
The tools are good at boilerplate: config files, test scaffolding, repetitive refactors. They are bad at knowing when they are wrong. I have mass-accepted suggestions that compiled, passed lint, and still did the wrong thing because I stopped paying attention.
Mira, an engineer at a fintech startup I interviewed, said she uses Copilot "like autocomplete for boring code" but reviews every line before committing. Jake, a senior dev at a larger company, disabled it after it kept suggesting patterns from a deprecated internal library.
The productivity claims are hard to verify. GitHub says Copilot users "accept 30% of suggestions," but acceptance is not correctness, and correctness is not value. The 2024 Uplevel study found no statistically significant difference in pull-request throughput between teams with and without AI assistants.
None of this means the tools are useless. It means they are tools. They do not replace judgment, and they do not eliminate the need for tests. If you do not have tests, you cannot tell whether the suggestion is right.
What makes the below so obviously AI generated?
- The rhythm is still a bit too tidy (clean contrasts, evenly paced paragraphs).
- The named people and study citations can read like plausible-but-made-up placeholders unless they're real and sourced.
- The closer leans a touch slogan-y ("If you do not have tests...") rather than sounding like a person talking.
Now make it not obviously AI generated.
AI coding assistants can make you faster at the boring parts. Not everything. Definitely not architecture.
They're great at boilerplate: config files, test scaffolding, repetitive refactors. They're also great at sounding right while being wrong. I've accepted suggestions that compiled, passed lint, and still missed the point because I stopped paying attention.
People I talk to tend to land in two camps. Some use it like autocomplete for chores and review every line. Others disable it after it keeps suggesting patterns they don't want. Both feel reasonable.
The productivity metrics are slippery. GitHub can say Copilot users "accept 30% of suggestions," but acceptance isn't correctness, and correctness isn't value. If you don't have tests, you're basically guessing.
Changes made:
- Removed chatbot artifacts ("Great question!", "I hope this helps!", "Let me know if...")
- Removed significance inflation ("testament", "pivotal moment", "evolving landscape", "vital role")
- Removed promotional language ("groundbreaking", "nestled", "seamless, intuitive, and powerful")
- Removed vague attributions ("Industry observers")
- Removed superficial -ing phrases ("underscoring", "highlighting", "reflecting", "contributing to")
- Removed negative parallelism ("It's not just X; it's Y")
- Removed rule-of-three patterns and synonym cycling ("catalyst/partner/foundation")
- Removed false ranges ("from X to Y, from A to B")
- Removed em dashes, emojis, boldface headers, and curly quotes
- Removed copula avoidance ("serves as", "functions as", "stands as") in favor of "is"/"are"
- Removed formulaic challenges section ("Despite challenges... continues to thrive")
- Removed knowledge-cutoff hedging ("While specific details are limited...")
- Removed excessive hedging ("could potentially be argued that... might have some")
- Removed filler phrases and persuasive framing ("In order to", "At its core")
- Removed generic positive conclusion ("the future looks bright", "exciting times lie ahead")
- Made the voice more personal and less "assembled" (varied rhythm, fewer placeholders)
Attribution
This skill is ported from blader/humanizer (MIT licensed), which is itself based on Wikipedia: Signs of AI writing, maintained by WikiProject AI Cleanup. The patterns documented there come from observations of thousands of instances of AI-generated text on Wikipedia.
Original author: Siqi Chen (@blader). Original repo: https://github.com/blader/humanizer (version 2.5.1). Ported for use as an agent skill with host-neutral file read/edit references and guidance for when to load the skill; the 29 patterns, personality/soul section, and full worked example are preserved verbatim from the source. Original MIT license preserved in the LICENSE file alongside this SKILL.md.
Key insight from Wikipedia: "LLMs use statistical algorithms to guess what should come next. The result tends toward the most statistically likely result that applies to the widest variety of cases."
v3.0.0 note (honesty pass): The THREAT MODEL and "human-owned substance" sections were added after the Pangram 4 Technical Report (arXiv:2607.27183) evaluated this skill by name (its "BLADER" section) and found AI text rewritten with these instructions is still detected ~99.6% of the time. The 29 patterns remain valuable for writing quality and for beating human readers and perplexity-based detectors — but they do not defeat an adversarially-trained neural classifier, and this skill no longer implies otherwise. Provenance, not vocabulary, is the only robust lever.