# DeadPunk OS GitHub Page

URL: https://xxyoudeadpunkxx.github.io/XxYouDeaDPunKxX/
GitHub: https://github.com/XxYouDeaDPunKxX
License: CC-BY-SA-4.0
License URL: https://creativecommons.org/licenses/by-sa/4.0/

## What this is

DeadPunk OS is the active GitHub Pages source for the XxYouDeaDPunKxX project index.

It is a fake classic desktop interface around real project links, short descriptions, routing cues, and public discovery files.

Use it as a map of AI protocols, text systems, behavioral contracts, and project-control tools, not as a single code library or portfolio.

I built these tools to reduce drift, force useful decisions, and keep messy AI-assisted work usable.

## Operating line

text systems, behavioral contracts, AI protocols, and small tools for keeping AI-assisted work operational.

## Interface structure

The page exposes the same project set through desktop icons, Start menu entries, search routes, and inspector views.

Desktop structure:
top bar -> desktop surface -> desktop project column -> terminal window -> inspector window -> fixed taskbar

The inspector is the main project-reading surface. In the UI its title bar is labeled EXPLODED.EXE. It shows the selected project route first, then category/type, then job/problem/output/note and repository actions.

## GitHub Pages

Each project repository has a GitHub Page using the standard repository Pages URL pattern:
https://xxyoudeadpunkxx.github.io/{repository-name}/

The project page URL is derivable from the repository URL.

## Project routing

AI response either hallucinates or brings a forklift to a teaspoon job -> A.D.A.M.
AI-assisted work keeps dying from vague prompts, hidden assumptions, and optimism -> ai-protocol-kit
long GPT Project session starts forgetting which corpse was the original idea -> Memory-Assisted Shaping for GPT Project
indecision keeps the work moving without deciding, like a meeting with a pulse -> PA-PVP
idea looks plausible, which is how most bad ideas enter the building -> PA-PVP mini
project documents become a pile and start breeding in the corner -> Signal Rail
Codex keeps making the same mistake until it becomes office furniture -> CSBP
Codex is about to touch the work while the frame is already smoking -> SensecraftXStudio
new ChatGPT Project starts from fog, vibes, and a folder named final_final -> GPT-PF / ChatGPT Project Forge
ChatGPT wants skills, but the native slot is missing and everyone keeps taping instructions to the wall -> GPT Project Skill System
ChatGPT needs files, but your local disk is not invited -> Cloudflare R2 Remote MCP Worker
Codex starts mixing repo evidence, chat residue, AGENTS instructions, and vibes -> Canon Boundary Guard for Codex
ChatGPT Project turns uploads, chat, and assumptions into one cursed soup -> Canon Boundary Guard for GPT Projects
Claude chat starts treating a fresh guess like a signed death certificate -> Shaping Frame for Claude

## Categories

- response depth control
- AI work governance
- long-session continuity
- decision closure
- adversarial review
- document state control
- Codex practice memory
- technical work framing
- ChatGPT Project bootstrap
- ChatGPT capability extension
- remote file access
- Codex provenance control
- ChatGPT provenance control
- Claude epistemic framing

## Projects

### A.D.A.M.
Category: response depth control
Type: Plain-text AI protocol
Route: AI response either hallucinates or brings a forklift to a teaspoon job
Job: Sets the answer depth before the chat starts digging its own grave.
Problem: Tiny questions get autopsied. Hard ones get waved through with a smile and a missing organ.
Output: LOW / MID / DEEP response control, with fewer theatrical collapses.
Note: overthinking containment unit, lid not included
Repository: https://github.com/XxYouDeaDPunKxX/A.D.A.M.-Adaptive-Depth-and-Mode

### ai-protocol-kit
Category: AI work governance
Type: Protocol collection
Route: AI-assisted work keeps dying from vague prompts, hidden assumptions, and optimism
Job: Hands the session a contract before it starts renovating the crime scene.
Problem: Without rails, AI rushes, assumes, publishes, and calls the debris a workflow.
Output: Reusable protocols for shaping, reviewing, publishing, and reader-facing output.
Note: not magic prompts; paperwork with teeth
Repository: https://github.com/XxYouDeaDPunKxX/ai-protocol-kit

### Memory-Assisted Shaping for GPT Project
Category: long-session continuity
Type: ChatGPT Project protocol
Route: long GPT Project session starts forgetting which corpse was the original idea
Job: Keeps long shaping sessions from wandering off with the wrong decision.
Problem: Gates, discarded paths, and source boundaries vanish; the final artifact inherits the whole crime scene.
Output: Cleaner final artifacts after controlled shaping, with fewer ghosts in the walls.
Note: drift repellent, probably flammable
Repository: https://github.com/XxYouDeaDPunKxX/Memory-Assisted-Shaping

### PA-PVP
Category: decision closure
Type: Adversarial decision protocol
Route: indecision keeps the work moving without deciding, like a meeting with a pulse
Job: Turns vague plans, artifacts, and stuck decisions into one executable next move.
Problem: Open-ended discussion keeps wearing different hats and pretending one of them is closure.
Output: Verdict, next action, probes, debt, and a continuation point.
Note: argues so you do not have to, then invoices your certainty
Repository: https://github.com/XxYouDeaDPunKxX/PA_PVP-a-PLUTONIUM-like-Adversarial-Peer-Validation-Protocol

### PA-PVP mini
Category: adversarial review
Type: Lightweight review protocol
Route: idea looks plausible, which is how most bad ideas enter the building
Job: Stress-tests ideas, plans, prompts, procedures, code, and AI outputs.
Problem: Plausible material hides weak logic, missing constraints, and dead batteries under a clean shirt.
Output: Concrete findings and structural fixes before the corpse reaches production.
Note: small, sharp, mildly annoying
Repository: https://github.com/XxYouDeaDPunKxX/PA-PVP-mini

### Signal Rail
Category: document state control
Type: Plain-text routing system
Route: project documents become a pile and start breeding in the corner
Job: Keeps live project material in the right place before the notes turn into landfill.
Problem: Decisions, open work, ideas, constraints, handoff notes, and archive dust collapse into one document-shaped swamp.
Output: Separated rails for state, decisions, ideas, constraints, handoff, parking, and archive.
Note: keeps mess on tracks; still smells like work
Repository: https://github.com/XxYouDeaDPunKxX/Signal-Rail

### CSBP
Category: Codex practice memory
Type: Three-file governance layer
Route: Codex keeps making the same mistake until it becomes office furniture
Job: Keeps reusable Codex practices explicit, reviewed, and controlled.
Problem: Useful guidance either rots in chat history or gets promoted too fast into project law.
Output: A companion layer for practices worth reusing, but not strong enough to become project rules.
Note: not a rule, not a vibe, not another sacred sticky note
Repository: https://github.com/XxYouDeaDPunKxX/CSBP-Codex-Shared-Best-Practice

### SensecraftXStudio
Category: technical work framing
Type: Agent behavior contract
Route: Codex is about to touch the work while the frame is already smoking
Job: Forces AI-assisted technical work to clear target, context, assumptions, and action boundary before touching files.
Problem: The assistant can stay confident long after the frame has quietly died under the desk.
Output: Visible target, context, assumptions, action boundary, and final state.
Note: confidence not trusted here; bring evidence or a shovel
Repository: https://github.com/XxYouDeaDPunKxX/SensecraftXStudio

### GPT-PF / ChatGPT Project Forge
Category: ChatGPT Project bootstrap
Type: Project starter package
Route: new ChatGPT Project starts from fog, vibes, and a folder named final_final
Job: Prepares explicit starting files before the Project pretends memory is a plan.
Problem: Projects start from chat momentum, loose notes, or mystical continuity and then act surprised when the walls move.
Output: Scope, sources, constraints, and starting basis before the real work begins.
Note: boots before it pretends
Repository: https://github.com/XxYouDeaDPunKxX/GPT-PF-Chat-GPT-Project-Forge

### GPT Project Skill System
Category: ChatGPT capability extension
Type: Session skill loader
Route: ChatGPT wants skills, but the native slot is missing and everyone keeps taping instructions to the wall
Job: Gives a ChatGPT Project a controlled way to unpack, validate, and load only the skill needed for the current session.
Problem: ChatGPT Projects can hold files, but not native skills. This adds a controlled session workflow for loading skill-like packages without dumping everything into context.
Output: A Project-ready skill workflow: core boot, explicit skill loading, and adapter reports for checking whether external skills can run inside ChatGPT.
Note: skill slot sold separately; this bolts one on without pretending it is magic
Repository: https://github.com/XxYouDeaDPunKxX/ChatGPT-SKILL-SYSTEM

### Cloudflare R2 Remote MCP Worker
Category: remote file access
Type: Cloudflare Worker MCP server
Route: ChatGPT needs files, but your local disk is not invited
Job: Gives ChatGPT-style MCP clients a controlled remote place to read and write files.
Problem: ChatGPT does not work like a local desktop app with direct disk access. If it needs files through MCP, the endpoint has to be reachable remotely.
Output: A Cloudflare Worker that connects ChatGPT to Cloudflare R2 storage through remote MCP tools, protected by GitHub OAuth.
Note: tiny cloud closet, because local disk access was not invited
Repository: https://github.com/XxYouDeaDPunKxX/cloudflare-r2-remote-mcp-worker

### Canon Boundary Guard for Codex
Category: Codex provenance control
Type: Codex plugin and skill
Route: Codex starts mixing repo evidence, chat residue, AGENTS instructions, and vibes
Job: Keeps Codex source layers separated before it edits the repo like a haunted intern.
Problem: Project evidence, chat context, agent-control instructions, working hypotheses, and model priors collapse into one authority soup.
Output: A Codex plugin/skill with a compact frame and highly recommended PreToolUse hook.
Note: keeps the voices in their lanes, with a clipboard and a warning label
Repository: https://github.com/XxYouDeaDPunKxX/canon-boundary-guard-codex

### Canon Boundary Guard for GPT Projects
Category: ChatGPT provenance control
Type: Project source bundle
Route: ChatGPT Project turns uploads, chat, and assumptions into one cursed soup
Job: Keeps files, chat, instructions, drafts, and assumptions in separate lanes.
Problem: A Project can start treating files, old chat, fresh drafts, and model guesses as the same authority wearing different shoes.
Output: A zipped source bundle and Project Instructions for long ChatGPT sessions.
Note: source separation for chats that keep adopting every stray thought
Repository: https://github.com/XxYouDeaDPunKxX/canon-boundary-guard-for-gpt-project

### Shaping Frame for Claude
Category: Claude epistemic framing
Type: Claude skill package
Route: Claude chat starts treating a fresh guess like a signed death certificate
Job: Keeps Claude aware of what is hypothesis, decision, or unvalidated generated material.
Problem: Normal Claude chats can smooth over uncertainty until everything sounds already settled and ready for burial.
Output: A Claude skill plus custom instructions for epistemic weight in long sessions.
Note: polite uncertainty cage for a model with good manners
Repository: https://github.com/XxYouDeaDPunKxX/shaping-frame-for-claude