dispatching-parallel-agents

name	dispatching-parallel-agents
description	Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies

Dispatching Parallel Agents

Overview

When you have multiple unrelated failures (different test files, different subsystems, different bugs), investigating them sequentially wastes time. Each investigation is independent and can happen in parallel.

Core principle: Dispatch one agent per independent problem domain. Let them work concurrently.

Teams Availability Check

Before dispatching, check if the Teams feature is available. Teams provides coordinated parallel execution with inter-agent communication, shared task lists, and lifecycle management.

digraph teams_check {
    "2+ independent tasks?" [shape=diamond];
    "TeamCreate tool available?" [shape=diamond];
    "4+ tasks?" [shape=diamond];
    "Team Mode" [shape=box style=filled fillcolor=lightgreen];
    "Simple Parallel Dispatch (Task tool)" [shape=box];
    "Sequential or single agent" [shape=box];

    "2+ independent tasks?" -> "TeamCreate tool available?" [label="yes"];
    "2+ independent tasks?" -> "Sequential or single agent" [label="no"];
    "TeamCreate tool available?" -> "4+ tasks?" [label="yes"];
    "TeamCreate tool available?" -> "Simple Parallel Dispatch (Task tool)" [label="no - fallback"];
    "4+ tasks?" -> "Team Mode" [label="yes"];
    "4+ tasks?" -> "Simple Parallel Dispatch (Task tool)" [label="2-3 tasks, simpler"];
}

How to check: Attempt to use TeamCreate. If it's not in your available tools, fall back to simple parallel dispatch with the Task tool.

When to prefer Team Mode even with 2-3 tasks:

• Tasks may spawn follow-up work
• Agents need to communicate findings to each other
• You want progress tracking via shared task list
• Tasks have dependency ordering (some must finish before others start)

When to Use

digraph when_to_use {
    "Multiple failures?" [shape=diamond];
    "Are they independent?" [shape=diamond];
    "Single agent investigates all" [shape=box];
    "One agent per problem domain" [shape=box];
    "Can they work in parallel?" [shape=diamond];
    "Sequential agents" [shape=box];
    "Parallel dispatch" [shape=box];

    "Multiple failures?" -> "Are they independent?" [label="yes"];
    "Are they independent?" -> "Single agent investigates all" [label="no - related"];
    "Are they independent?" -> "Can they work in parallel?" [label="yes"];
    "Can they work in parallel?" -> "Parallel dispatch" [label="yes"];
    "Can they work in parallel?" -> "Sequential agents" [label="no - shared state"];
}

Use when:

• 2+ tasks with different root causes or domains
• Multiple subsystems broken independently
• Each problem can be understood without context from others
• No shared state between investigations

Don't use when:

• Failures are related (fix one might fix others)
• Need to understand full system state
• Agents would interfere with each other (editing same files)

Simple Parallel Dispatch (Fallback / 2-3 Tasks)

When Teams isn't available or you have only 2-3 independent tasks:

1. Identify Independent Domains

Group failures by what's broken:

• File A tests: Tool approval flow
• File B tests: Batch completion behavior
• File C tests: Abort functionality

2. Create Focused Agent Tasks

Each agent gets:

• Specific scope: One test file or subsystem
• Clear goal: Make these tests pass
• Constraints: Don't change other code
• Expected output: Summary of what you found and fixed

3. Dispatch in Parallel

Task("Fix agent-tool-abort.test.ts failures")
Task("Fix batch-completion-behavior.test.ts failures")
Task("Fix tool-approval-race-conditions.test.ts failures")
// All three run concurrently

4. Review and Integrate

When agents return:

• Read each summary
• Verify fixes don't conflict
• Run full test suite
• Integrate all changes

Team Mode (4+ Tasks or Need Coordination)

When Teams is available and you have many independent tasks or need inter-agent communication:

1. Create Team and Task List

TeamCreate(team_name="fix-failures", description="Fix 6 test failures across 4 files")

2. Create Tasks with Dependencies

TaskCreate(subject="Fix abort tests", description="Fix 3 failures in agent-tool-abort.test.ts...")
TaskCreate(subject="Fix batch tests", description="Fix 2 failures in batch-completion.test.ts...")
TaskCreate(subject="Fix race conditions", description="Fix 1 failure in tool-approval.test.ts...")
TaskCreate(subject="Integration verification", description="Run full suite after all fixes land")

// Task 4 depends on 1-3 completing
TaskUpdate(taskId="4", addBlockedBy=["1", "2", "3"])

3. Spawn Teammates

Task(name="fixer-1", team_name="fix-failures", subagent_type="general-purpose",
     prompt="You are a teammate. Check TaskList, claim an available task, fix it, mark complete.")
Task(name="fixer-2", team_name="fix-failures", subagent_type="general-purpose",
     prompt="You are a teammate. Check TaskList, claim an available task, fix it, mark complete.")
Task(name="fixer-3", team_name="fix-failures", subagent_type="general-purpose",
     prompt="You are a teammate. Check TaskList, claim an available task, fix it, mark complete.")

4. Monitor and Coordinate

As team lead:

• Messages from teammates arrive automatically
• Use SendMessage to answer questions or provide context
• Use TaskList to check progress
• When blocked tasks unblock (dependencies complete), assign or let teammates claim them

5. Shutdown and Cleanup

When all tasks are complete:

SendMessage(type="shutdown_request", recipient="fixer-1", content="All done")
SendMessage(type="shutdown_request", recipient="fixer-2", content="All done")
SendMessage(type="shutdown_request", recipient="fixer-3", content="All done")
// After all acknowledge:
TeamDelete()

Team Mode Advantages Over Simple Dispatch

Feature	Simple Dispatch	Team Mode
Communication	None (fire-and-forget)	DMs, broadcasts
Progress tracking	Wait for all to return	Live task list
Dependencies	None	blocks/blockedBy
Follow-up work	Manual re-dispatch	Teammate claims next task
Error handling	Read final report	Teammate messages you mid-task
Lifecycle	Implicit (agent exits)	Explicit shutdown protocol

Agent Prompt Structure

Good agent prompts are:

1. Focused - One clear problem domain
2. Self-contained - All context needed to understand the problem
3. Specific about output - What should the agent return?

Fix the 3 failing tests in src/agents/agent-tool-abort.test.ts:

1. "should abort tool with partial output capture" - expects 'interrupted at' in message
2. "should handle mixed completed and aborted tools" - fast tool aborted instead of completed
3. "should properly track pendingToolCount" - expects 3 results but gets 0

These are timing/race condition issues. Your task:

1. Read the test file and understand what each test verifies
2. Identify root cause - timing issues or actual bugs?
3. Fix by:
   - Replacing arbitrary timeouts with event-based waiting
   - Fixing bugs in abort implementation if found
   - Adjusting test expectations if testing changed behavior

Do NOT just increase timeouts - find the real issue.

Return: Summary of what you found and what you fixed.

Common Mistakes

❌ Too broad: "Fix all the tests" - agent gets lost ✅ Specific: "Fix agent-tool-abort.test.ts" - focused scope

❌ No context: "Fix the race condition" - agent doesn't know where ✅ Context: Paste the error messages and test names

❌ No constraints: Agent might refactor everything ✅ Constraints: "Do NOT change production code" or "Fix tests only"

❌ Vague output: "Fix it" - you don't know what changed ✅ Specific: "Return summary of root cause and changes"

❌ Too many teammates: Spawning 8 agents for 4 tasks ✅ Right-sized: Match teammate count to task count (max 4-5 teammates)

When NOT to Use

Related failures: Fixing one might fix others - investigate together first Need full context: Understanding requires seeing entire system Exploratory debugging: You don't know what's broken yet Shared state: Agents would interfere (editing same files, using same resources)

Verification

After agents return (either mode):

1. Review each summary - Understand what changed
2. Check for conflicts - Did agents edit same code?
3. Run full suite - Verify all fixes work together
4. Spot check - Agents can make systematic errors