Notes on running agent fleets.

What we learn building Maxxwell — supervising many coding agents at once, and what actually breaks when you do.

---

• Autonomous agent runs vs supervised loops 2026-10-06

  When you’re touching architecture or security, the real choice isn’t “which agent model”. It’s: do you let it run autonomously end-to-end, gremor-style, or do.

• Orchestration as a product vs a codebase 2026-10-05

  You have two real options for multi-agent coding orchestration:

• Who should schedule agent work, and at what layer 2026-10-04

  If you live inside The Cog already, use its native orchestration for a single project or tightly scoped workflow. Once you’re supervising more than one cog.

• Autonomous planners vs human-led agent fleets 2026-10-03

  Short answer: if you want a single autonomous "brain" that decomposes work, dispatches agents, and merges results with minimal human input, you reach for a.

• When to run one coding agent vs many 2026-10-02

  If you already live in Claude Code, Cursor, or Codex, the real choice is simple:

• Focus cues vs inbox hacks for agent fleets 2026-10-01

  Past a couple of coding agents, the hard part isn’t getting output. It’s staying on top of what’s going on without spending your whole day in Slack, email, and.

• Herd view vs tmux for supervising agents 2026-09-30

  You already know how to run more agents. The problem is supervising them without spending your whole day in tab management and half-read transcripts.

• Measuring agent throughput: herded vs unmanaged 2026-09-29

  Past a couple of coding agents, your bottleneck stops being tokens and starts being you. You’re answering questions, tracking stalls, and cleaning up rework.

• Shell scripts vs a manager for agent fleets 2026-09-28

  Past three or four coding-agent sessions, your problem stops being model quality and starts being you. You’re answering prompts, hopping tmux panes, and.

• Boards vs terminals for herding coding agents 2026-09-27

  You already know the problem: the agents are fine; you are the bottleneck. You’re juggling Claude Code, Codex, Cursor agents, and your day turns into answering.

• Instrumenting Agent Fleets with OpenTelemetry 2026-09-26

  Most teams adopt agents one developer at a time. The real pain starts when several people have fleets running and nobody can see what’s actually happening.

• A safe, repeatable AI refactoring workflow 2026-09-25

  AI coding IDEs are good enough now that the problem isn’t “can it write code?”-it’s “can I trust what it changed across 40 files while I was in a meeting?”

• Logging AI coding agents into Git and CI 2026-09-24

  You can run eight agent sessions just fine. You become the bottleneck when you can’t see which one is doing what, which commit came from which session, or why.

• Your First Week Running Multiple Coding Agents 2026-09-22

  Past two or three agent sessions, the model stops being the bottleneck. You do.

• Policy checks and guardrails for AI coding agents 2026-09-21

  You already know your AI coding agents can write code. The hard part is stopping them from merging unsafe or non-compliant changes while you’re juggling ten.

• Run four coding agents on one feature 2026-09-20

  Most people hit the same wall with coding agents: the model is fast, but you’re stuck babysitting six terminals. The feature gets built, but only if you.

• Keeping multiple coding agents on one plan 2026-09-19

  You don’t hit the wall because agents write bad code. You hit it because three of them quietly stopped sharing the same picture of the repo.

• Turning a feature spec into parallel agent tasks 2026-09-18

  Past a certain point you stop waiting on the models and start waiting on yourself.

• Sandboxes vs Control Planes 2026-09-18

  AI coding tools crossed the novelty line a while ago. If you’re reading this, the problem isn’t “does Claude / Copilot help?” - it’s “I have eight sessions.

• Workflow layers vs infrastructure-first stacks 2026-09-18

  Developers are past “does AI coding work?” and into “how do I run five agents without becoming the bottleneck?”. The market split is now clear.

• Orchestrating coding agents vs automating jobs 2026-09-18

  AI coding assistants are mainstream now, but once you run more than a couple, your attention becomes the bottleneck.

• Human approval gates for AI coding agents 2026-09-18

  AI coding agents are now fast enough that your attention, not their output, is the bottleneck. The trick is adding human-in-the-loop checkpoints that catch.

• Designing an AI Coding Agent Orchestration Plan 2026-09-18

  AI coding agents are now table stakes, but multi-agent setups mostly fail in the same boring ways: unclear goals, no responsibility boundaries, and no checkpoints. You get more output, not more progress.

• Agents vs scripts in your CI pipeline 2026-09-12

  You can wire AI into CI today. The hard part isn’t the API call, it’s deciding where agents belong and where a dumb, deterministic script is still the right.

• Running Claude Code, Codex and Cursor together 2026-09-11

  Most of us are past the "one prompt, one answer" phase. The problem now is coordinating multiple long-running coding agents without turning yourself into a.

• When a tmux script stops being enough 2026-09-10

  You can get surprisingly far coordinating coding agents with tmux and some bash. Then one day you realize the real bottleneck is you.

• How to tell which coding agent is stuck 2026-09-09

  Past two or three coding agents, you stop being limited by model quality and start being limited by your own attention. You have more terminals than you can.

• Orchestrating agents: greenfield vs legacy 2026-09-09

  Most teams now have more AI output than attention. The real problem is how to run and coordinate multiple AI coding agents without breaking prod or wasting.

• Monitoring and Debugging an AI Agent Swarm 2026-09-08

  When you're running more than a couple of AI coding agents, the bottleneck is no longer model output. It's your ability to see what's happening and intervene.

• Triage boards vs live agent status streams 2026-09-08

  You’re already running multiple coding agents in parallel. The new bottleneck isn’t tokens, it’s your attention. The question is: do you spend that attention.

• Wiring Coding Agents Across Repos and Monorepos 2026-09-08

  Most people hit the same ceiling the moment they run agents on more than one repo: the models are fine, but you are now the scheduler for ten terminals.

• Agent-native IDEs vs AI plugins vs orchestrators 2026-09-08

  AI coding tools have stopped being “one assistant in your IDE” and turned into multi-agent systems. The output is fine; the problem is keeping your head.

• Central conductor vs peer-to-peer agents 2026-09-01

  Past one or two coding agents, the problem stops being model quality and becomes orchestration. You’re suddenly a traffic controller for parallel work that can.

• Orchestrated agents vs codegen pipelines 2026-08-31

  Developers are now the bottleneck, not the models. Once you’re running five or ten coding agents, the hard part is keeping them aimed, not getting more code.

• Running Claude Code and Codex in parallel 2026-08-30

  Most people who run Claude Code or Codex seriously hit the same wall: the agents scale, your attention doesn’t.

• Running Claude Code agents in parallel 2026-08-29

  Why the first thing that breaks when you run several coding agents on one repo is the working directory, and what git worktrees fix about it.

• Single agents vs orchestrated fleets 2026-08-29

  Past one or two coding agents, the problem stops being code generation and starts being you babysitting terminals.

• IDE plugins vs. agent orchestration layers 2026-08-29

  Most teams already run AI coding tools. Stack Overflow’s 2025 survey says 51% of professional developers use AI daily, and JetBrains reports 85% regularly use.