Skip to main content

Concurrent coding agents

Firebender offers simple primitives to do this:
IsolationWhen to use
TabsNone (same directory)Lightweight and good for changes in completely different files, or for read-only operations like exploring the codebase
Parallel agentsFull (worktrees)Independent features/PRs that can have conflicts and you don’t want affecting your cwd. Auto heals from conflicts from the configured base
SubagentsManaged by AIComplex tasks where you want Firebender to delegate focused work in separate contexts
If done well, your overall throughput of good changes increases dramatically: Zone of productivity.

Common pitfalls

Parallelization overhead

Problem: Conflicts, higher error rates, and context-switching costs can eat into productivity gains. Solution:
  • Use Parallel agents for isolated changes that auto-heal merge conflicts from base
  • Use Subagents when you want the main agent to delegate focused work in parallel
  • Prefer smaller, narrowly-scoped tasks: they have the highest merge rate and lowest context-switching cost
  • Subagents: create focused callable agents for operational tasks like verification, PR review, or doc sync
Focused subagents scoped to specific tasks dramatically improve accuracy. Create your own with /agent or see Subagents.

Staying in flow

Problem: Managing AI agents can feel like being an engineering manager: frequent context switches, waiting on slow responses, and forgetting your original intent (“doorway effect”). Solution:
  • Write, Ask, and Plan modes: switch between implementation, read-only exploration, and planning as the task evolves
  • Plan mode: AI researches approaches and asks clarifying questions before coding
  • GLM 4.7 and GPT-5.2: fast agentic models for quick iteration
  • Commands: quickly insert pre-written prompts or task descriptions you use often
Use /help to quickly create a command based on a previous chat. See Commands.