LinkedIn's Smoke Detector

Victor Grycuk Traicovich

Victor Grycuk Traicovich

— 5 min read

So, I got tired of seeing Vendedores de Humo on LinkedIn, the affiliate link preachers, hustle prophets, and emoji-obsessed lunatics posting 14 times a week about how their morning coffee manifested a 7-figure pipeline.

LinkedIn loves that type of content. Engagement means revenue.
But it sucks for those of us who just want news and networking.

So I built a tool that detects them, and satirizes the heck out of them.

Now when I see one of them, I don’t groan any more.
I created a Smoke Detector.

Disclaimer: This is in constant development, and can be changed at any given time!

What is the Smoke Detector?

The Smoke Detector analyzes public LinkedIn profiles and turns them into collectible “specimens.” It doesn’t just say how smoky a profile is, it builds a playful, coherent character around it based on their characteristics.

  • Humo Score (0–100)
  • Satirical classifications (e.g., Professionals, Humo Vendors, LinkedIn Lunatics)
  • Six behavioral stats (competence, credibility, perception management, commercial intent, attention seeking, content volume)
  • Achievements, evolution stages, battle dynamics, habitat, and a deep‑dive analysis
0:00
/0:29

How it works (in brief)

  • I use a custom made web browser plugin on any public profile that looks interesting to capture important information
  • The profile’s public text is analyzed by a chain of AI agents (via n8n)
  • Results are saved to a lightweight database (PocketBase)
  • The site renders a new card and the collection updates

The website

The main collection lives at the website, which it is a responsive grid of specimen cards with flip animations, search/filter controls, and detailed modal views. Each card shows the Humo Score, rarity, stats, achievements, and a snippet of the analysis. Click any card to see the full breakdown: behavioral reasoning, natural habitat, evolution chain, battle dynamics, and combat abilities.

0:00
/1:03

The browser plugin

A Chrome extension that only activates on LinkedIn profile pages. When opened, it displays clickable sections (About, Experience, Education, etc.) that correspond to different parts of the profile. Click a section to navigate there automatically and extract the text. Once it has captured the sections I want, it sends everything to the n8n analysis pipeline with one click.

After processing completes, opening the plugin again shows the specimen card data. It's a seamless way to collect and analyze profiles without leaving LinkedIn.

Capturing information
A fully analysed specimen

Collecting Cards

Each collected profile appears as a card on the site. The grid is responsive, cards flip on hover, and you can open a detail modal for the deep dive. If I'm viewing a profile I’ve already collected, the card info appears directly on the plugin's popup.

Card anatomy

  • Humo Score + rarity badge
  • Category and type (e.g., LinkedIn Lunatic)
  • Six stats with tooltips
  • Achievements
  • Their professional analisis
  • Quote, natural habitat, and behaviour
  • And more.

The six stats

Each profile is analized and scored with 6 stats that represent their virtue and wickedness:

🎯
Competence: Measures the individual's demonstrated professional capabilities, technical knowledge, and ability to deliver quality work. Higher scores indicate proven expertise and reliable performance in their field.
🛡️
Credibility: Evaluates the person's reliability, consistency, and trustworthiness in professional settings. This includes their reputation, track record of keeping commitments, and overall professional integrity.
👁️
Perception Management: Assesses how well someone curates and manages their professional image, personal brand, and public perception. This includes strategic self-presentation and narrative control.
💲
Commercial Intent: Measures the degree to which someone's professional activities are driven by commercial goals, sales objectives, or business development opportunities.
📢
Attention Seeking: Evaluates behaviours aimed at gaining attention, visibility, or engagement from others. This can include controversial takes, frequent posting, or dramatic storytelling for engagement.
🧠
Content Volume: Measures the amount and frequency of content someone produces across professional platforms. High volume can indicate either valuable thought leadership or potential spam behaviour.

Professionals are driven by competence/credibility, Humo Vendors by perception/commercial intent, and LinkedIn Lunatics by attention/content volume.

The Humo Score

The Humo Score is basically:

“How much of a LinkedIn Fake are you?”

And it ranges from 0 to 100.

But it’s not a flat score, it adapts based on the user type. A salesperson spamming automation tools daily has a different “humo ceiling” than a burned-out developer posting memes about frameworks. How the humo score is exactly calculated depends on the type.

In simple terms: professionals get “less humo” the more competent/credible they are; vendors get “more humo” the more image‑driven they are; lunatics blend brand‑craft with sheer volume and attention tactics.

The tone system (satire that adapts)

The AI doesn’t write in one voice for everyone. It selects a tone based on the dominant behaviors, and scales the intensity.

  • Professional (nature documentary): respectful, curious, evidence‑based
  • Satirical (surgical deconstruction): irony that highlights image over substance
  • Cosmic Horror (for extreme cases): algorithmic apocalypse meets content deluge

Combination logic:

  • Professional + Humo: respectful disappointment
  • Professional + Lunatic: confused admiration
  • Humo + Lunatic: satirical horror

Tone applies consistently across habitat, evolution, attacks, and quotes; so each “specimen” feels coherent, not random.

A game‑inspired professional analysis

  • Natural Habitat + Behaviour: Environments that fit observed behavior (e.g., “conference corridors”, “comment‑section amphitheaters”)
  • Evolution Chain: entry form → current state → likely next evolution
  • Battle Dynamics: a rock‑paper‑scissors loop
    • Professional beats Humo Vendor (competence defeats perception)
    • Humo Vendor beats LinkedIn Lunatic (focused marketing beats chaos)
    • LinkedIn Lunatic beats Professional (overwhelming volume disrupts competence)
  • Combat Abilities: regular, special, status, and an ultimate; all derived from behavioral patterns
  • Achievements: 50+ badges that adjust stats/rarity (e.g., “Title Collector”, “Spam Overlord”, “Master Manipulator”)
  • Naming & Quotes: fictionalized, behavior‑driven, never identifying real people

GDPR & privacy

This is satire, not shaming. It’s built with guardrails.

  • What’s NOT stored: email, phone, DMs, profile photos, usernames, direct LinkedIn links, or any private info
  • What IS analyzed (but not stored): public titles, about sections, and posts; the stored output is AI‑generated, paraphrased characterizations
  • An additional agent scrubs potentially identifying details before saving
  • Takedown: if you believe your public content has been unintentionally represented, email [email protected]. Basic verification (e.g., a profile link) may be requested.

This tool is meant to be humorous and respectful. It is designed with safeguards to avoid storing or misusing personal information in any way.

Why I built it

  • To make the feed bearable, and sometimes delightful
  • To explore a transparent, ethics‑first satire engine
  • To show how lightweight AI + automation can craft playful, coherent experiences

What’s next

  • More achievements and smarter contradiction detection
  • Better tooltips and accessibility polish
  • A deeper technical write‑up (n8n workflow, prompts, stat maths, and rarity algorithm)

Read more