How to run multiple agents simultaneously — across sessions, within a session, and across complex workflows — to multiply your team's throughput without multiplying chaos.
EPISODES
The mental model for parallel agent development — why sequential is the wrong default, and how to think in concurrent workstreams.
Ep 2How to structure work for parallel Claude Code sessions: worktrees, focused prompts, rotation cadence, and PR-per-session strategy.
Ep 3The Agent tool deep-dive: fan-out patterns, passing context, collecting results, and avoiding runaway subagent cost spirals.
Ep 4Patterns for agents in different sessions that need to share state — worktrees, shared memory files, and merge orchestration.
Ep 5Building always-on engineering automation: scheduled agents, hooks pipelines, and multi-step workflow definitions.
Ep 6Permissions, approval gates, spend guardrails, and observability for teams running agents at scale across multiple projects.
BonusAgent Teams with mailbox + task board, dynamic workflows spawning hundreds of parallel subagents, the GAN pattern, and cost controls at swarm scale.
Teaching Claude your craft — Skills, Context Engineering, MCP & tools, and rolling it all out across your team.
Start Season 2 →