Field guides, comparisons, and ideas for builders running AI coding agents at portfolio scale.
Command Fleet's orchestration layer is agent-agnostic, so a local Qwen model becomes a first-class agent on the same board as Claude Code, Codex, and Gemini. How to route the right work to it.
A coding agent that runs entirely on your machine — local Qwen models via Ollama, no API key, fully offline. The strengths, the honest limits, and the setup.
Vibe Kanban is winding down. Here's what to look for in a replacement — and why a portfolio-scale, local-first orchestrator is the natural next step.
A practical workflow for founders and agencies juggling a portfolio of apps — workspaces, parallel runs, and review queues.
Cloud agents are convenient until your client's source code lives on someone else's servers. The case for keeping code, keys, and data on your machine.
You don't have to pick one. A pragmatic guide to matching each coding agent to the task — and switching per job.
How Command Fleet turns one sentence into a planned, built, reviewed, merged, and deployed app — and where you stay in control.
Building is the easy part; shipping is where demos fall over. Stack packs, one preview/deploy manifest, and credential-gated deploys to six platforms.
Agents finish work fast; that doesn't make it right. Read the diff, run a verify gate, and work a six-point checklist before you merge.
AI agents are powerful and occasionally careless. A practical security guide to keeping API keys and secrets out of agent prompts.
Subscription vs API, the hidden markup of bundled tools, and why bring-your-own plus per-task model routing keeps the bill down.
The unlock for parallel AI agents isn't a bigger model — it's git worktrees. Branch isolation, safe concurrency, and clean integration.
An AI editor and an agent orchestrator solve different problems. When to reach for each — and why you'd use both.
Want an autonomous AI engineer without the black box? Go local and agent-agnostic, with a review gate and no per-run markup.
The task is the program. The anatomy of a good agent task — acceptance criteria, scope, and a reusable template.
Three things make agents productive from the first task: a setup script, secrets in the vault, and a stack pack.
Local doesn't mean manual. The honest tradeoffs of local-first vs cloud AI agents — and when each one wins.
"Where does our code go?" should be one sentence. Keep client source local, secrets out of prompts, clients separated.
A plain-English primer: what an AI coding agent is, what it does, and how it differs from autocomplete.
Run many agents at once: the conductor pattern, isolation and scheduling, and how results get reviewed and merged.
A board is a state machine; a chat is a transcript. Why a Kanban board beats a chat window for running agents.
One prompt can't plan or recover. How a dependency-aware task graph lets agents run in parallel — and bounce back from failure.
You can't filter your way to safety. Defend agents with least privilege, secrets out of context, and a small blast radius.
The most reliable control over agents is what you don't grant. Scopes, sandboxes, and worktrees, explained.
Don't review code that doesn't build. A verify gate runs your build and tests before you ever look.
Review is a conversation. How to give feedback that makes the next agent run better — and capture it as criteria.
Fast, autonomous authors are safe behind gates. CI/CD for agents: verify gates, preview, promotion, and rollbacks.
Tests prove code runs; a preview proves the product works. See an AI-built app running before you ship it.
Bundled tools price in a markup on every run. Bring-your-own pays the provider once — here's the math.
Nine habits that cut the bill without slowing down: route by task, scope small, bring your own, measure per feature.
For a fleet of agents, the right strategy is the boring one: one task, one short branch, merged in order.
Parallel agents don't have to mean merge hell. Design conflicts down with isolation, merge order, and small tasks.
What Windsurf does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What GitHub Copilot does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Replit does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Bolt.new does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Lovable does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What v0 does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Cline does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Aider does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Continue does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Sourcegraph Cody does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Augment Code does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Amazon Q Developer does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What JetBrains AI does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Tabnine does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Qodo does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What CodeRabbit does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Greptile does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Factory.ai does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What OpenHands does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Google Jules does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What OpenAI Codex does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Trae does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Zed does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Conductor does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Terragon does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Tabby does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Create.xyz does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Databutton does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Goose does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
What Warp does well — and where Command Fleet adds a portfolio board, your choice of agent, an autonomous build loop, and built-in deploys.
