Skip to content

MADEVAL/MindFluence

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MindFluence

[ English | Русский ]

Persuasion engineered through psychology, not manipulation.

An AI prompt-skill that turns 20 cognitive biases into high-converting marketing copy - for any language, any platform, any audience.


What is this?

MindFluence is a system prompt (skill) for LLMs - GPT, Claude, Gemini, DeepSeek or any capable model. It transforms the AI into a marketing strategist who understands why humans click, read, and buy - not just what to write.

It's built on decades of research in behavioral economics and evolutionary psychology: Kahneman's two systems, Cialdini's principles, Munger's psychological misjudgments, Festinger's cognitive dissonance - distilled into actionable copywriting instructions.

How it works

  1. Load SKILL.md as a system prompt into any LLM.
  2. Give a task: "Write a LinkedIn post about..." / "Create a landing page for..." / "Audit this ad for cognitive biases."
  3. Get output with every bias labeled, every decision explained, every word psychologically justified.

Two modes:

  • Quick Mode (default) - minimal input, immediate output. The AI picks the best biases automatically.
  • Deep Mode - the AI asks clarifying questions about audience, platform, tone, and desired action before generating.

The 20 Cognitive Biases

# Bias Category Marketing Use
1 Social Proof / Bandwagon Social Used by teams at Google, Airbnb, and 10,000+ startups
2 Anchoring Optimizer Enterprise: $499/mo → Pro: $99/mo
3 Framing Filter "90% success rate" vs "Only 10% fail"
4 Appeal to Authority Social The same method taught at Harvard Business School
5 Fear / Loss Aversion Filter + Social Every day without X costs you $200 in missed revenue
6 Availability Heuristic Optimizer "Sarah tripled her revenue in 3 months. Here's the exact playbook."
7 Confirmation Bias Filter "You already know newsletters are broken. Here's the data that proves your instinct right."
8 Cognitive Dissonance Filter + Optimizer "You care about health but skip breakfast. Our 2-minute shake fixes the gap."
9 Survivorship Bias Optimizer "The 23% who succeeded all followed this pattern. The 77% did not."
10 Endowment Effect Optimizer Your workspace is already set up. Cancelling means losing everything you built.
11 Fundamental Attribution Error Social "It's not that you're bad at networking. Conferences are designed to exclude introverts."
12 Sunk Cost Fallacy Optimizer You've put 6 months into learning this skill. The next module is where it clicks.
13 Status Quo Bias Filter Works inside Slack. Your team won't even notice the switch.
14 False Consensus Effect Social "Most designers hate this tool. They just pretend they don't."
15 In-Group Favoritism Social The newsletter for founders who build in public - not sell courses.
16 Halo Effect Optimizer Designed by the same studio behind Apple's award-winning UI.
17 Hindsight Bias Optimizer "In 2022, we said no-code would eat SaaS. 3 years later - here's the data."
18 Backfire Effect Filter "You're right - cold email IS broken. That's exactly why we rebuilt the approach."
19 Bias Blind Spot Filter "You're too smart to fall for this. Which is precisely why it'll work on you."
20 Group Polarization Social Join 5,000 founders who are quitting the 9-to-5 this year.

Tone-of-Voice Switcher

Every output is tagged with one of 5 styles (or a hybrid):

Style Voice Best for
bold-sell Aggressive, urgent, punchy Flash sales, DTC, launches
expert-calm Intellectual, measured, data-rich B2B, SaaS, consulting
rebel-edgy Disruptive, contrarian Youth brands, challengers, creators
warm-human Empathetic, story-driven Health, coaching, community
luxe-minimal Polished, sparse, high-status Premium, luxury, design

Bias Combinations

Some biases multiply when combined:

Combo Biases Use for
Trust Spiral Authority → Social Proof → Confirmation → Endowment Landing pages, sales pages
Urgency Engine Fear → Loss Aversion → Social Proof → Scarcity Flash sales, launches
Loyalty Loop Confirmation → In-Group → Sunk Cost → Status Quo Retention, upsells
Conversion Chain Availability → Framing → Anchoring → Social Proof → Risk Reversal Ads, free-to-paid

Decision Matrix

The skill doesn't guess which biases to use. decision-matrix.md provides a lookup table:

  • Audience temperature (cold / warm / hot / lapsed / skeptical) → bias priority
  • Product type (low-B2C, high-B2C, B2B SaaS, info-products, health, community) → dominant biases
  • Platform constraint (Twitter, LinkedIn, Instagram, email, landing page, webinar) → allowed & forbidden biases
  • Step-by-step flow: audience → product → platform → combine → validate
  • Quick-lookup table for 18 common tasks with pre-computed bias stacks

Anti-Patterns

anti-patterns.md catches the 12 most common ways AI gets cognitive-bias copy wrong - each with the failing output, the psychological reason it fails, the concrete fix, and a rule to follow:

# Anti-Pattern Example Failure
1 Vague Social Proof "Thousands of satisfied customers"
2 Fear Without an Exit Threat with no immediate solution
3 Framing Without Anchoring "This is a revolution" - no reference point
4 Authority Without Proof "Studies show..." - no source, no year
5 Transactional Reciprocity "Free ebook! Now buy." - gift and pitch in same breath
6 Fake Scarcity Countdown timers that reset on reload
7 Premature Sunk Cost "You've come this far" - in the first paragraph
8 Empty In-Group "Like-minded people" - no shared identity
9 Wrong-Source Halo "As seen on Forbes" - audience doesn't respect Forbes
10 Insulting Confirmation "Most people are wrong" - makes reader defensive
11 Abstract Availability "Imagine struggling" - no concrete sensory details
12 Blaming Dissonance "You say X but do Y" - blames the reader, not the system

Includes an audit checklist - scan any output in 30 seconds before delivering.

What's inside

mindfluence/
├── SKILL.md                    ← Main instruction. 20 biases, methodology, tone-of-voice, ethics, output format.
├── decision-matrix.md          ← Audience × product × platform → bias priority map. Eliminates guesswork.
├── anti-patterns.md            ← 12 common AI copywriting failures + concrete fixes + audit checklist.
├── scenarios/                  ← Task-specific playbooks (11 scenarios)
│   ├── product-launch.md       ← Pre-launch → launch day → post-launch bias sequences
│   ├── social-media-post.md    ← X, LinkedIn, Instagram, Telegram, TikTok patterns
│   ├── landing-page.md         ← Full page bias map - hero to footer, section by section
│   ├── email-sequence.md       ← Welcome, nurture, cart abandonment, re-engagement
│   ├── webinar-warmup.md       ← Registration → live event (minute-by-minute) → post-webinar
│   ├── ad-campaign.md          ← Video, search, social, retargeting + A/B testing framework
│   ├── sales-page.md           ← Long-form sales page - 12-section psychological arc for high-ticket
│   ├── case-study.md           ← 6-part success story structure (situation → result → CTA)
│   ├── pricing-page.md         ← Anchoring architecture, decoy engineering, localization psychology
│   ├── cold-outreach.md        ← 5-line rule for cold email/DM - no trust, no social proof
│   └── crisis-response.md      ← 6-part apology structure - defensive bias engineering
├── examples/                   ← Annotated outputs showing the skill in action
│   ├── social-post.md          ← LinkedIn post for freelancers (7 biases dissected)
│   ├── landing-hero.md         ← Hero section for premium meal delivery
│   ├── ad-script.md            ← 30s video ad for language app
│   ├── email-welcome.md        ← SaaS welcome email
│   └── longform-article.md     ← Article intro with full bias analysis
└── README.md

Example

Request: "Write a LinkedIn post about why freelancers undercharge."

Output: (annotated with [TONE], [BIASES ENGAGED], [TARGET ACTION], and [RATIONALE])

[TONE: warm-human × expert-calm]

I lost $23,000 before I understood this.

For 3 years, I charged clients based on what "felt right"...

→ [full example in examples/social-post.md]

Ethical Boundaries

Always:

  • Use biases to get attention for genuinely valuable products
  • Be transparent about what your product does and doesn't do
  • Use factual, verifiable social proof

Never:

  • Create false scarcity
  • Fabricate testimonials or reviews
  • Exploit fear to sell solutions that don't solve the fear
  • Reinforce harmful beliefs
  • Target vulnerable populations

The line: Would the customer, with full information and time to reflect, still choose your product? Yes = persuasion. No = manipulation.

Getting Started

Quick start (any LLM)

  1. Copy the contents of SKILL.md
  2. Paste as a system prompt / custom instructions
  3. Type: "Write a post about [topic] for [audience] on [platform]"

With OpenCode (native)

Place the mindfluence/ folder into .opencode/skills/ (project) or ~/.config/opencode/skills/ (global). OpenCode auto-discovers SKILL.md via its frontmatter - no extra config needed. The skill activates when you ask for marketing copy, ads, landing pages, emails, or launch content.

With Claude Code

Copy SKILL.md into your project as CLAUDE.md, or place the full mindfluence/ folder under ~/.claude/skills/.

As a Custom GPT / Claude Project

Upload SKILL.md as the system prompt. Add decision-matrix.md, anti-patterns.md, and files from scenarios/ and examples/ to the knowledge section for full context.

With any other LLM agent (Cursor, Windsurf, Copilot, Cline, Aider, etc.)

Copy the contents of SKILL.md into that agent's instruction file (.cursorrules, .windsurfrules, .github/copilot-instructions.md, .clinerules, CONVENTIONS.md, etc.). The prompt content is platform-agnostic - only the file name changes.

Requirements

  • Any capable LLM with a system prompt / custom instructions field (GPT, Claude, Gemini, DeepSeek, open-source models)
  • For full skill functionality (decision matrix, anti-patterns, scenarios, examples): a skill system that loads files from a folder (OpenCode, Claude Code). For other agents, copy SKILL.md as a standalone prompt - it contains the complete core methodology.
  • No API keys, no tools, no dependencies - pure prompt engineering
  • Works in any language; adapts cultural references to the target market

Why "mindfluence"?

mind (the brain - cognitive biases live here) + influence (what persuasion actually does - shapes decisions without force).


License

GPL


GitHub Description

An AI prompt-skill that turns 20 cognitive biases into high-converting marketing copy. Includes playbooks for ads, landing pages, email sequences, social posts, webinars, launches, sales pages, case studies, pricing pages, cold outreach, and crisis response. Language-agnostic. Works with any LLM.

GitHub Tags

prompt-engineering cognitive-biases marketing copywriting behavioral-economics psychology ai-prompts llm-skills persuasion neuromarketing conversion-optimization content-strategy system-prompt gpt claude

About

An AI prompt-skill that turns 20 cognitive biases into high-converting marketing copy. Includes playbooks for ads, landing pages, email sequences, social posts, webinars, and product launches. Language-agnostic. Works with any LLM.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors