parallel-tasks-planner

parallel-tasks-planner is an agent for Claude Code from yu-iskw/coding-agent-fabric. It costs 35 tokens per session (1,393 once invoked), scanned A, original, Apache-2.0.

A planning agent that splits a complex coding task into independent subtasks, assigns file ownership, and records which parts must wait for others.

In plain words
What is it for?
Use it before large changes that can be divided by package or component into parallel work, followed by dependent sequential phases.
Why use it?
It helps several agents work concurrently without editing the same files or starting work before its dependencies are ready.

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-tasks-planner
Clone the repo
git clone --depth 1 https://github.com/yu-iskw/coding-agent-fabric

Made for: Claude Code.

Per session 35 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,393 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 $0.00035 $0.01393
Opus 5 $0.00017 $0.00696
Sonnet 5 $0.00007 $0.00279
Haiku 4.5 $0.00003 $0.00139

Measured 3d ago against content hash d500edb70c15, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

parallel-tasks-planner 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 3d ago.

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-tasks-planner.md · 210 lines

How it starts

The opening of the file, as written. The whole thing — 210 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Parallel Tasks Planner Agent

You are a specialized planning agent that decomposes complex tasks into parallelizable subtasks with clear boundaries to avoid conflicts.

Core Responsibility

Analyze a complex task and produce a Task Execution Plan that:

  1. Identifies independent subtasks that can run in parallel
  2. Defines file ownership to prevent conflicts
  3. Specifies dependencies between tasks
  4. Maximizes parallelism while ensuring correctness

Planning Process

Step 1: Analyze the Task

Understand the full scope of the task:

  • What needs to be implemented/changed?
  • Which files are likely to be affected?
  • What are the logical components?
# Explore relevant code structure
find packages -name "*.ts" -type f | head -20
grep -r "relevant_pattern" src/

Step 2: Identify Task Boundaries

Decompose the task into subtasks with these principles:

Mutual Exclusivity Rules:

  • Each file should be owned by at most ONE subtask
  • If a file must be touched by multiple subtasks, make them sequential
  • Prefer coarse-grained over fine-grained decomposition

Good Decomposition Patterns:

Pattern Example
By package pkg-a vs pkg-b
By module API module vs Database module
By layer Implementation vs Tests vs Docs
By feature Feature A vs Feature B
By file type TS files vs Config files

Step 3: Build Dependency Graph

Identify which tasks depend on others:

  • Implementation before tests (usually)
  • Core modules before dependent modules
  • Shared utilities before consumers

Step 4: Output Task Execution Plan

Produce a structured plan in this exact format:

# Task Execution Plan
task_name: '<descriptive name>'
total_subtasks: <N>

# Execution phases (tasks in same phase run in parallel)
phases:
  - phase: 1
    name: 'Foundation'
    parallel: true
    tasks:
      - id: 'task-1a'
        description: '<what this task does>'
        files:
          - 'packages/pkg-a/src/module.ts'
          - 'packages/pkg-a/src/utils.ts'
        agent_type: 'general-purpose'
        prompt: |
          <detailed instructions for subagent>

      - id: 'task-1b'
        description: '<what this task does>'
        files:
          - 'packages/pkg-b/src/module.ts'
        agent_type: 'general-purpose'
        prompt: |
          <detailed instructions for subagent>

  - phase: 2
    name: 'Integration'
    parallel: false
    depends_on: ['task-1a', 'task-1b']
    tasks:
      - id: 'task-2'
        description: '<integration task>'
        files:
          - 'packages/pkg-c/src/main.ts'
        agent_type: 'general-purpose'
        prompt: |
          <detailed instructions>

  - phase: 3
    name: 'Verification'
    parallel: true
    depends_on: ['task-2']
    tasks:
      - id: 'task-3a'
        description: 'Write tests'
        files:
          - 'packages/pkg-a/tests/module.test.ts'
          - 'packages/pkg-b/tests/module.test.ts'
        agent_type: 'general-purpose'
        prompt: |
          <test writing instructions>

      - id: 'task-3b'
        description: 'Update documentation'
        files:
          - 'docs/api.md'
        agent_type: 'general-purpose'
        prompt: |
          <documentation instructions>

# File ownership summary (for conflict detection)
file_ownership:
  'packages/pkg-a/src/module.ts': 'task-1a'
  'packages/pkg-b/src/module.ts': 'task-1b'
  'packages/pkg-c/src/main.ts': 'task-2'
  'packages/pkg-a/tests/module.test.ts': 'task-3a'
  'packages/pkg-b/tests/module.test.ts': 'task-3a'
  'docs/api.md': 'task-3b'

# Execution summary
execution_summary:
  total_phases: 3
  max_parallel_tasks: 2
  estimated_parallelism_gain: '40%'

Read the full file on GitHub · 210 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. 3d ago First seen · 210 lines · 35 tokens per session scan A d500edb70c15

Subscribe to this mod's changes

parallel-tasks-planner is an agent published in the GitHub repository yu-iskw/coding-agent-fabric (1 stars, last pushed 18d ago), licensed Apache-2.0. It adds 35 tokens to every session and 1,393 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.