One dev runs a 4-agent team that researches, reviews, and publishes automatically
_jaydeepkarale · x · 2026-08-21
The author assigned four AI agents distinct roles—research, review, polish, publish—to automate his research-to-publish workflow: one finds backend engineering resources, one reviews GitHub MRs, one does final review, and one posts finished work to Discord. They can browse the web, edit repos, create MRs, review each other, hand off work, and run the routine every morning. He describes it as hiring a tiny team rather than using an AI assistant.
More from coding & agent
- Patronus Open Sources 200+ Hours of Real Figma Design Trajectories — Div_pradeep · 2026-08-21
- Open-source x402-cleanweb-agent saves 80% tokens by cleaning web content — EstablishmentTough18 · 2026-08-21
- Ex-Linktree CRM lead open-sources Orbit: 130 MCP tools for Claude to gate marketing emails — CleanPucks · 2026-08-21
- Model vs Harness: Why Great Models Fail in Poor Agents — arpit_bhayani · 2026-08-21
- Editing Video with Codex: Defining workflow with four key questions — gabrielchua · 2026-08-21
- Codebases for the AI Era: Hard Constraints and Type Systems Over Style Guides — cedric_chee · 2026-08-21