Alex Fazio is an AI engineer at AI Garden who builds agent-development tools and open-source examples. His work includes Plankton, which applies code-quality checks as an agent edits files, and a collection of CrewAI tutorials.[1][2][3]
Start with Plankton →
X Contributions
Feedback at the moment of an edit
Plankton uses Claude Code hooks to run configured checks when an agent writes code. The checks can cover style, types, security and complexity, returning failures as feedback before the change reaches commit or review. The repository marks it as a research project under active development, with compatibility depending on the supported hook interface.[2]
Examples for learning agent orchestration
Fazio’s CrewAI Quickstart gathers notebooks and recipes demonstrating agent workflows and tools. His broader profile describes work on context and harness engineering across several agent SDKs. These materials give developers implementations to inspect and adapt rather than relying only on high-level descriptions of multi-agent systems.[1][3]
Ideas
Its documentation explains setup, language-specific tools and the effects on existing code. An edited file can expose pre-existing issues as well as newly introduced ones. This makes the project a concrete example of moving verification earlier in an agent workflow, while also requiring the operator to understand which checks and hooks are installed.[2]
Tools, writing & talks
Tools
Start here
Plankton
github.comWrite-time code-quality checks and feedback built around Claude Code hooks.
Start with the supported-hook and existing-codebase behavior, then inspect the configured checks. Trying it requires the documented Claude Code version and local language tools.
- ARCgenticagithub.com
Agentic system that solves ARC-AGI-2 puzzles by having sub-agents write and test Python transform programs.
Catalog workOriginal source - CrewAI Quickstartgithub.com
Notebooks and recipes for learning CrewAI workflows and integrations.
Creator and repository owner[3]