CITATION.cff

cff-version: 1.2.0
message: "If you use AIGX or its benchmark, please cite it as below."
title: "AIGX: AI Genome Exchange - A Benchmark-Validated Context Format for AI Coding Agents"
abstract: >-
  AIGX (AI Genome Exchange) is an open context format for AI coding agents that stores a codebase's rules
  in a centralized .aigx/ directory with a per-file boundary index, injecting nothing into source code. In
  a controlled benchmark holding the rule set constant and varying only the context format, AIGX ranked
  first on mean score, pass@1, and hidden-test pass rate on both Claude Haiku 4.5 and Claude Sonnet 4.6 at
  n=60, and survived roughly two dozen deliberate attempts to beat it. The top formats are statistically
  tied on the composite mean; AIGX's defensible edge is cross-model consistency, robustness under
  challenge, simplicity, and being the only format validated in a controlled benchmark.
type: software
authors:
  - family-names: Parisotto
    given-names: Grégory
    email: gregory@feex.it
keywords:
  - AI coding agents
  - context engineering
  - context format
  - LLM
  - agent instructions
  - code generation
  - developer tools
  - benchmark
  - agentic AI
  - prompt engineering
  - software development
license: MIT
version: "1.1"
date-released: "2026-06-15"
url: "https://github.com/Lolner95/AIGX"
repository-code: "https://github.com/Lolner95/AIGX"
preferred-citation:
  type: software
  title: "AIGX: AI Genome Exchange - A Benchmark-Validated Context Format for AI Coding Agents"
  authors:
    - family-names: Parisotto
      given-names: Grégory
  year: 2026
  url: "https://github.com/Lolner95/AIGX"