How does it optimize the task execution order?

The skill analyzes dependencies between deliverables to create a graph, identifying which tasks must be sequential and which can be executed in parallel groups to save time.

When does the skill decide to aggregate or split tasks?

It aggregates tasks that share the same domain and profile if the file count is low (<10). It splits tasks if they involve mixed execution modes or exceed 15 files to maintain AI context focus.

How are skills assigned to the generated tasks?

Skills are resolved by querying the project's architecture metadata based on the specific module and profile (like implementation or testing) associated with each deliverable.

What is the primary function of the Phase Refine Plan skill?

It acts as a translation layer that takes deliverables from a solution outline and turns them into structured, executable tasks that include specific skill requirements and file paths.

Can this skill detect errors in my project plan?

Yes, it specifically checks for circular dependencies and missing modules in the architecture, preventing the execution of an invalid or incomplete plan.

Phase Refine Task Planner

Name: Phase Refine Task Planner
Author: cuioss

bycuioss

Productivity & Workflow

Transforms architectural solution outlines into optimized, executable tasks for AI coding assistants.

About

The Phase Refine Task Planner is a core orchestration skill for Plan Marshall that bridges the gap between design and execution. It processes solution deliverables to create granular, executable tasks while automatically resolving domain-specific skills from the project architecture. By applying intelligent aggregation and splitting rules, it optimizes the workflow for maximum efficiency and parallel execution, ensuring that Claude Code receives clear, structured instructions that comply with strict task contracts.

Key Features

Parallelism Calculation: Computes optimal execution order and groups tasks that can be performed simultaneously.
0 GitHub stars
1:N Task Mapping: Automatically generates separate implementation and testing tasks for single deliverables.
Intelligent Aggregation: Merges related small deliverables to minimize overhead while maintaining focus.
Dependency Validation: Constructs dependency graphs and detects circular references to prevent execution errors.
Skill Resolution: Dynamically queries project architecture to assign required domain skills to specific tasks.

Use Cases

Optimizing AI developer workflows by grouping small, related code changes.
Converting high-level architectural designs into a sequence of actionable coding tasks.
Ensuring consistent skill application across different modules in a large-scale project.

About

Key Features

Parallelism Calculation: Computes optimal execution order and groups tasks that can be performed simultaneously.
0 GitHub stars
1:N Task Mapping: Automatically generates separate implementation and testing tasks for single deliverables.
Intelligent Aggregation: Merges related small deliverables to minimize overhead while maintaining focus.
Dependency Validation: Constructs dependency graphs and detects circular references to prevent execution errors.
Skill Resolution: Dynamically queries project architecture to assign required domain skills to specific tasks.

Use Cases

Optimizing AI developer workflows by grouping small, related code changes.
Converting high-level architectural designs into a sequence of actionable coding tasks.
Ensuring consistent skill application across different modules in a large-scale project.