I Taught My Second Brain to Run Multi-Agent Coding Workflows (Live Session)

I Taught My Second Brain to Run Multi-Agent Coding Workflows (Live Session)

More

Descriptions:

Cole Medin hosts a live coding session previewing Archon, his open-source workflow engine for AI-assisted software development, ahead of its planned public release. Archon’s core goal is to make AI coding deterministic and repeatable — packaging multi-step agent pipelines (planning, implementation, code review, evaluation) into structured, executable workflows. Medin describes it as “N8N but for software development,” drawing a comparison to Stripe’s internal “Minions” system, which combines deterministic and non-deterministic nodes in a similar fashion.

During the session, Medin walks through a GitHub issue resolution workflow he uses dozens of times daily, showing how Archon orchestrates Claude Code across multiple roles: a builder agent, an evaluator that judges output against sprint criteria and writes to a feedback file, and a lightweight reporter using Claude Haiku for token efficiency. A “second brain” layer sits above these workflows, routing tasks across projects and maintaining state in JSON files. The system is designed to run workflows in parallel — the session culminates in a stress test where multiple issues are dispatched simultaneously.

Archon has been available privately within Medin’s Dynamis community and is slated for public release the following Wednesday, with a full onboarding-focused livestream the Saturday after. The session is aimed at developers already comfortable with Claude Code and agent concepts who want to see how production-level AI coding pipelines handle complex, multi-phase software tasks reliably at scale.


📺 Source: Cole Medin · Published April 02, 2026
🏷️ Format: Livestream

1 Item

Channels

1 Item

People