freshcrate
Skin:/
Home > Frameworks > agent-arch

agent-arch

No description

Why this rank:Release freshnessStrong adoptionHealthy release cadence

Description

README

Agent Architecture Book

ะ ัƒััะบะฐั ะฒะตั€ัะธั Contributing guide Code of Conduct

Publishable documentation site and book about modern secure AI agent architecture.

Agent Architecture Book Preview

What this repository contains

  • a GitHub Pages site built with MkDocs and Material for MkDocs
  • a multilingual documentation structure
  • the first published part of the book on safe agent architecture
  • a small runnable reference package in agent_runtime_ref/
  • a modern Python-first tooling setup based on uv

Local development

uv sync --group docs --group dev
uv run mkdocs serve

The local site will be available at http://127.0.0.1:8000/.

Checks

uv run ruff check .
uv run ty check
uv run pytest --cov=agent_runtime_ref --cov-report=term-missing
uv run mkdocs build --strict

Reference package

The repository includes a minimal runnable skeleton package:

.venv/bin/python -m agent_runtime_ref

It gives you a compact reference runtime for the book:

  • runtime and policy layer
  • capability catalog and approved inventory
  • memory path, telemetry, approvals, and rollout checks
  • lifecycle-oriented artifacts for change records, artifact bundles, and retirement plans
  • YAML-driven configs for the operational skeleton

Quick examples:

.venv/bin/python -m agent_runtime_ref simulate-run
.venv/bin/python -m agent_runtime_ref inspect-agent
.venv/bin/python -m agent_runtime_ref inspect-lifecycle
.venv/bin/python -m agent_runtime_ref inspect-session
.venv/bin/python -m agent_runtime_ref export-session --output artifacts/session-demo-001.json
.venv/bin/python -m agent_runtime_ref export-eval-dataset --output artifacts/eval-dataset.json

The canonical walkthrough, full CLI list, and config overview live on the docs page:

Optional research dependencies

If you want notebooks or data analysis tooling:

uv sync --group research

The research group already includes marimo and polars.

Publishing

The repository includes a GitHub Actions workflow for GitHub Pages:

  • build with uv
  • strict mkdocs build --strict
  • Pages deployment from the docs-prod branch

First GitHub Pages setup

actions/configure-pages@v5 has an important limitation: if Pages have never been enabled in the repository before, the default GITHUB_TOKEN may not be able to bootstrap the Pages site automatically.

There are two correct ways to handle the first setup:

  1. Enable Pages once in Settings -> Pages and choose GitHub Actions.
  2. Add a PAGES_PAT secret with sufficient permissions, and the workflow can enable Pages automatically.

If the github-pages environment has deployment branch restrictions, make sure it explicitly allows deployments from docs-prod.

PAGES_PAT should be a real token, not GITHUB_TOKEN:

  • for a Personal Access Token: repo or Pages write permission
  • for a GitHub App: administration:write and pages:write

Branch model

  • main is the source-of-truth development branch
  • docs-prod is the publication branch used for GitHub Pages deployment

Stack

  • uv for environment and dependency management
  • ruff for linting
  • ty for type checking
  • MkDocs + Material for MkDocs for publishing
  • Mermaid and Observable Plot for visual content

License

This repository is published under CC BY-SA 4.0.

Release History

VersionChangesUrgencyDate
main@2026-10-08Latest activity on main branchHigh10/8/2026
0.0.0No release found โ€” using repo HEADHigh4/11/2026

Dependencies & License Audit

Loading dependencies...

Similar Packages

SimpleLLMFuncA simple and well-tailored LLM application framework that enables you to seamlessly integrate LLM capabilities in the most "Code-Centric" manner. LLM As Function, Prompt As Code. ไธ€ไธช็ฎ€ๅ•็š„ๆฐๅˆฐv0.8.4
swarmsThe Enterprise-Grade Production-Ready Multi-Agent Orchestration Framework. Website: https://swarms.ai6.8.1
agent-frameworkA framework for building, orchestrating and deploying AI agents and multi-agent workflows with support for Python and .NET.dotnet-1.28.0
langroidHarness LLMs with Multi-Agent Programming0.69.21
adk-javaAn open-source, code-first Java toolkit for building, evaluating, and deploying sophisticated AI agents with flexibility and control.v1.14.0

More in Frameworks

fastapiFastAPI framework, high performance, easy to learn, fast to code, ready for production
crewAIFramework for orchestrating role-playing, autonomous AI agents. By fostering collaborative intelligence, CrewAI empowers agents to work together seamlessly, tackling complex tasks.
crewaiCutting-edge framework for orchestrating role-playing, autonomous AI agents. By fostering collaborative intelligence, CrewAI empowers agents to work together seamlessly, tackling complex tasks.
simBuild, deploy, and orchestrate AI agents. Sim is the central intelligence layer for your AI workforce.