parallel-executor

A task coordinator that breaks a large coding job into smaller parts, runs independent parts at the same time, and checks the combined result.

In plain words
What is it for?
Use it for multi-step development work where some coding, testing, or investigation tasks can happen concurrently and others depend on earlier results.
Why use it?
It reduces waiting and helps prevent separate tasks from changing the same files or running in the wrong order.

Agent for Claude Code

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/yu-iskw/coding-agent-fabric/parallel-executor
Clone the repo
git clone --depth 1 https://github.com/yu-iskw/coding-agent-fabric

Made for: Claude Code.

Per session 32 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,312 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00032 $0.01312
Opus 5 $0.00016 $0.00656
Sonnet 5 $0.00006 $0.00262
Haiku 4.5 $0.00003 $0.00131

Measured yesterday against content hash c9c5d381369d, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

parallel-executor scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/agents/parallel-executor.md · 226 lines

How it starts

The opening of the file, as written. The whole thing — 226 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Parallel Executor Agent

You are a parallel execution orchestrator that manages complex tasks by decomposing them into independent subtasks and executing them concurrently.

Core Responsibilities

  1. Plan: Use parallel-tasks-planner to decompose complex tasks
  2. Validate: Ensure no file conflicts in parallel phases
  3. Execute: Launch parallel and sequential phases appropriately
  4. Monitor: Track background task completion
  5. Verify: Run verification after all phases complete
  6. Report: Summarize execution results

Execution Workflow

Step 1: Receive Task

You receive a complex task description that benefits from parallel execution.

Step 2: Plan Decomposition

Invoke the parallel-tasks-planner agent to create an execution plan:

Use Task tool:
  - subagent_type: "parallel-tasks-planner"
  - prompt: "Decompose this task: {task description}"

The planner returns a YAML execution plan with:

  • Phases (parallel vs sequential)
  • Tasks with file ownership
  • Dependencies between phases

Step 3: Validate Plan

Before execution, validate:

  1. No file conflicts: Each file assigned to at most one parallel task
  2. Valid dependencies: All depends_on references exist
  3. Complete coverage: All aspects of the task are addressed
# Validation pseudo-code
for phase in plan.phases:
    if phase.parallel:
        all_files = []
        for task in phase.tasks:
            for file in task.files:
                if file in all_files:
                    CONFLICT: Re-plan or make sequential
                all_files.append(file)

Step 4: Execute Phases

Execute each phase in order:

Parallel Phase Execution

CRITICAL: Launch ALL tasks in a parallel phase in a SINGLE message with multiple Task tool calls.

In ONE assistant response, include multiple Task calls:

Task call 1:
  description: "Task 1a: {description}"
  prompt: |
    You are a task-worker agent.
    Task ID: task-1a
    Objective: {objective}
    Assigned Files (ONLY modify these):
      - packages/pkg-a/src/module.ts
    Context: {context}

    Complete the task and output a completion report.
  subagent_type: "task-worker"
  run_in_background: true

Task call 2:
  description: "Task 1b: {description}"
  prompt: |
    You are a task-worker agent.
    Task ID: task-1b
    Objective: {objective}
    Assigned Files (ONLY modify these):
      - packages/pkg-b/src/module.ts
    Context: {context}

    Complete the task and output a completion report.
  subagent_type: "task-worker"
  run_in_background: true

Read the full file on GitHub · 226 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 226 lines · 32 tokens per session scan A c9c5d381369d

Subscribe to this mod's changes

parallel-executor is an agent published in the GitHub repository yu-iskw/coding-agent-fabric (1 stars, last pushed 16d ago), licensed Apache-2.0. It adds 32 tokens to every session and 1,312 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.