workflow

Describes the 'Gauntlet Loop', a viral 3-line Claude prompt that uses sub-agents with critics to iteratively improve output to a quality bar.

This article explains how to integrate DeepSeek into Codex via CC Switch, allowing you to use DeepSeek's models in Codex while retaining Codex's plugins and skills.

Explores ways to use git blame over time within GNU Emacs, offering practical techniques for navigating commit history.

The author introduces Orbit, a local-first workspace for managing multiple AI agents, using Markdown files for projects, tasks, decisions, and logs without requiring an account or cloud storage.

Aaron Levie shares insights on real-world AI agent adoption, arguing that agents are more like managing a process than chatting, and require workflow changes for big upside.

Yohei Nakajima describes upgrading his AI-agent stack to work hands-free from anywhere, with a chief of staff assigning tasks to project managers and builders, all coordinated via repos.

Tweet claiming an easy way to build AI automations, linking to a resource.

HAR is an open-source harness for orchestrating multi-agent coding workflows.

A tweet describing a personal knowledge management system where Claude reads and links markdown notes in a vault, acting as a librarian over a portable, model-agnostic second brain.

A researcher reflects on whether AI-assisted summarization has made them lazier or more capable, questioning if reliance on LLM outputs undermines genuine understanding while improving output quality.

A developer explains how to effectively use Claude Code subagents by capping and scoping them, rather than running many agents, to avoid context burn and overlapping work.

StepShot is an AI tool that converts real workflows into step-by-step guides.

Kiro Crew is an open-source AI agent platform, originally built inside Amazon, that coordinates engineering work across tools and runs tasks autonomously. It has been adopted by over 39,000 Amazon builders with nearly 500 contributors, and is now released publicly to let developers inspect, run, and extend it.

Yohei Nakajima set up a shared memory system across ChatGPT, Codex, Claude, and Claude Code, using local history for the coding tools and a scheduled recap skill for ChatGPT/Claude so any one can answer questions about recent work across all four.

Ray Fernando shares a workflow tip for running Codex agents more efficiently, claiming it avoids messy output and saves tokens, and is used by engineers at top FANG companies.

A developer shares experience building a customer support AI agent, highlighting that the human handoff and state management are the hardest parts, not the conversational ability.

Shubham Saboo discusses how graphs make agent organizations programmable, contrasting static workflow graphs with dynamic agent orgs that rewrite themselves during execution, and cites FarmTable as an example.

The author describes a workflow for running an AI coding agent overnight with a single goal, constraints, and a mandatory morning report, to maximize idle compute time.

A developer shares lessons from building a production n8n AI support workflow, emphasizing classification over generation, rule-based filters to cut costs, and grounding responses in a knowledge base to avoid hallucinated answers.

Introducing the open-source project dashi-taskboard, a local-first tool for managing Codex workflows via task kanban, supporting browser embedding and CLI, aimed at improving Vibe Coding efficiency.