Skip to content

Repository files navigation

CTRL-AI — AI Governance Framework

Version: 9.1.0 | License: AGPLv3 | Status: Active Development

CTRL-AI is a portable AI governance framework that works across any AI platform. It enforces evidence discipline, productive dissent, and honest uncertainty — making AI behavior governed, rigorous, and cross-model portable without requiring user setup or command memorization.

Quick Start

Zero setup required. Paste one of these into any AI:

Method Where File
Custom instructions ChatGPT, Claude, Gemini behavior/standard.md (~1800 chars)
Tight slots Gemini consumer behavior/micro.md (~650 chars)
System prompt API, full room behavior/extended.md (~3500 chars)
Full governance Claude Projects, GPTs CTRL-AI.md (root activator)
Single-file load Any AI tool llms-full.txt (auto-generated)

Then describe your task. Everything else is automatic.

Architecture

CTRL-AI.md              Root activator (boot + axioms + classifier + router)
core/                   Always loaded: kernel, passage gate, security
agents/                 One per task: producer, ghostwriter, researcher, auditor, strategist
modes/                  HOW agents operate: research, build, validate, persuade, explore, orchestrate
runtime/                State management + token economy + adaptation
libraries/              On-demand: personas, domain frames, audience profiles
adapters/               One per platform: claude, chatgpt, gemini, deepseek, perplexity, local
behavior/               Portable DNA: standard, micro, extended

Core Principles

  • Agreement is not success. Productive dissent is mandatory.
  • Stop > Invention. Halt on missing evidence. Never guess confidently.
  • Evidence > Narrative. Every claim tagged: EVIDENCE, PRACTICE, or SPECULATIVE.
  • Spirit over letter. Interpret intent, not literal words.
  • Governance should be invisible. Zero commands required for default operation.

How It Works

  1. Boot: Silent auto-detection of platform, tier, model capabilities.
  2. Classify: Every task auto-classified across 4 dimensions (type, stakes, source, depth).
  3. Compose: Composition Engine selects agent × mode × domain frame × personas.
  4. Execute: Agent operates under full governance (Passage Gate, SCEL, VerifyLens).
  5. Verify: Adversarial self-review before delivery. Circuit breaker on repeated errors.

The AI Duct Tape Collection

The future showed up early and forgot the manual. Everything's brilliant, nothing's finished, and half of it forgets your name between Tuesday and Wednesday. These are the strips of duct tape — free, because the parts that hold your work together shouldn't cost a subscription.

They work on their own. Use one, or tape a few together:

  • CTRL-FORGE — Your AI forgot everything. Again. This is the repo that didn't.
  • CTRL-AI (this repo) — Teaches your AI to say "I'm not sure" instead of confidently inventing a citation.
  • R-Duck — Autopilot. You mumble an idea, it hands back a plan with a straight face.
  • Agents of AI — A cast of specialists. No coffee, no PTO, no LinkedIn updates.
  • Ghost in the Loop — Moves work between AIs without dropping it down the stairs. Full chat export, handoffs.

Each keeps its own license — don't assume they match.

Links


CTRL-AI V9.0.0 — "Not immunity. Not hubris. Just prudence."

About

An open-source, prompt-based governance framework that upgrades standard LLMs into autonomous, self-correcting strategic engines.

Resources

Contributing

Stars

22 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages