Skip to content

Intent-Bound Authorization (IBA): Cryptographic protocol to prevent Intent Blindness & Confused Deputy in agentic AI. Includes 4-layer architecture, Semantic Firewall (<5ms), simulator, and ISO 42001 alignment.Intent-Bound Authorization (IBA): Making sure AI agents do what they're told, not just what they're capable of.

Notifications You must be signed in to change notification settings

Grokipaedia/iba-agentic-security

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 

Repository files navigation

Intent-Bound Authorization (IBA)

Cryptographic protocol that binds tool access to signed semantic intent — solving Intent Blindness and the Confused Deputy problem in agentic AI systems.

License: MIT Python Stars

Why IBA?

Legacy auth (OAuth, RBAC) gives agents a blank check once authorized. IBA enforces purpose fidelity at runtime with <5ms overhead — critical for enterprise agentic adoption in 2026.

Key Features

  • 4-layer architecture: Declaration → Binding → Validation → Enforcement
  • Semantic Firewall with ZKP privacy & drift detection
  • Prevents Confused Deputy, prompt injection drift, unauthorized escalation
  • Maps to ISO/IEC 42001, GDPR, HIPAA, etc.
  • Interactive simulator + reference implementation

Quick Start (Simulator)

git clone https://github.com/grokipaedia/iba-protocol.git
cd iba-protocol/simulator
pip install -r requirements.txt
python simulator.py

About

Intent-Bound Authorization (IBA): Cryptographic protocol to prevent Intent Blindness & Confused Deputy in agentic AI. Includes 4-layer architecture, Semantic Firewall (<5ms), simulator, and ISO 42001 alignment.Intent-Bound Authorization (IBA): Making sure AI agents do what they're told, not just what they're capable of.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages