AWS Builder Center

Builder Center

Artemisa: A Deterministic AI Agent Configurator Built with Kiro for the Kiro x Código Facilito Hackathon

Artemisa: A Deterministic AI Agent Configurator Built with Kiro for the Kiro x Código Facilito Hackathon

How we used Kiro to build Artemisa, a stateless, open-source generator of AI agent configuration bundles, during the Kiro x Código Facilito 2026 Hackathon.

Axel

Artemisa: A Hackathon Project Built with Kiro to Configure AI Agents

We Kiro was our primary AI coding assistant, and the project became a stateless, open-source generator of configuration bundles for AI agents.

What Artemisa Does

Artemisa turns 32 answers from a decision tree into a reproducible bundle of configuration files. The user answers through a web UI or API, reviews the output, and copies it into their project. No commands run, no files write automatically, and no LLM keys are required.
Each bundle includes:
  • artemisa.blueprint.json — every decision in one canonical file.
  • manifest.json — all artifacts with SHA-256 hashes.
  • docs/INSTALL.md and docs/WHY.md — how to apply the bundle and why it was built that way.
  • Target files for Cursor, Kiro, CodeRabbit, Kilo Code, and others, including steering rules, skills, MCP servers, RAG, and security policy.

Why It Is Different

Artemisa does not execute agents. It only generates configuration. This keeps the backend small and safe.
  • No runtime. The backend is pure generation, very cheap and tracable.
  • No state. No database, sessions, or files. Every request is a pure function of its body, so it scales horizontally with no shared storage.
  • Deterministic. Same answers and catalog version always produce the same bundle and hashes. Version mismatches return 409.

Built with Kiro

Kiro helped us design the architecture, write the TypeScript and Next.js code, shape the decision tree, and review the security model. It read the project context from AGENTS.md  and CONTEXT.md , so it stayed aligned with our constraints. The bundle also emits .kiro/ steering files, which Kiro can consume directly.

Try It

Artemisa is open-source under MPL-2.0 and runs as a container with no database or persistent disk.
  • Repository: https://github.com/VECTORG99/Artemisa
  • Live demo: https://artemisa-ai.netlify.app
# kiro# ai# open-source# kiro-agent# hackathon
Any opinions in this article are those of the individual author and may not reflect the opinions of AWS.
Enjoyed reading this content? Let the author know!

Your likes, comments, shares, and saves help creators reach more builders.

Loading recommendations

Loading article