How does Task Orchestration prevent context loss?

It uses a persistent MCP server to store task states, dependencies, and execution history, allowing Claude to resume work exactly where it left off in a new session.

Does it automatically update task statuses?

It detects 'cascade events' and recommends status changes based on your project's specific workflow configuration, ensuring consistent progress tracking.

What are specialist sub-agents?

Specialists are specific model configurations optimized for different tasks: Haiku for standard implementation, Sonnet for complex problem solving, and Opus for high-level architectural design.

Can this skill handle parallel tasks?

Yes, the skill automatically analyzes task dependencies to identify which tasks can be executed simultaneously in parallel batches to speed up development.

How does it optimize token usage?

It uses a Section Tag Taxonomy to filter and read only the specific parts of a task (like checklists or commands) that are necessary for the current operation, reducing token overhead by up to 60%.

Task Orchestration

Name: Task Orchestration
Author: jpicklyk

byjpicklyk

•

128

•

生产力与工作流

Manages complex coding workflows with dependency-aware parallel execution and automated specialist routing to eliminate context loss.

Task Orchestration is a high-performance skill designed to bridge the gap between architectural planning and implementation by managing granular development tasks. It enables Claude to handle multi-step features by identifying task dependencies, grouping them into parallel execution batches, and routing specific requirements to the most cost-effective specialist models (Haiku, Sonnet, or Opus). By leveraging a persistent MCP memory layer, it ensures that project state, progress, and technical context remain consistent across long development sessions, significantly reducing manual overhead and token consumption through intelligent tag filtering.

主要功能

01Persistent workflow state management to eliminate session context loss

02Token-optimized context reading using granular tag filtering

03Automated routing to specialist sub-agents (Implementation, Senior, Architect)

04128 GitHub stars

05Automatic dependency cascade detection and status progression

06Dependency-aware parallel task batching and execution

使用场景

01Managing long-running development sessions without losing project context

02Orchestrating specialized LLM roles for complex full-stack implementations

03Executing large-scale features with multiple interdependent sub-tasks

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jpicklyk/task-orchestrator task-orchestration

For use in Claude.ai and ChatGPT

Download Skill