Can this skill run multiple tasks at once?

Yes, it detects tasks marked with parallel hints in your PLAN.md and can launch and track multiple agents concurrently within a single response.

What happens if an implementation task gets blocked?

The skill enters a structured blocked state handler, presenting the specific reason for the blocker and offering options like retry, skip, abort, or manual assistance.

Does it automatically update my project documentation?

Yes, the skill is designed to update checkboxes in your PLAN.md file at the completion of each phase to ensure documentation matches the actual state of the codebase.

How does the Agent Coordination skill prevent context overload?

It uses an incremental loading protocol that only loads tasks for the current phase into TodoWrite, clearing completed tasks before moving to the next phase to keep the context focused.

Agentic Execution & Coordination

Name: Agentic Execution & Coordination
Author: rsmdt

byrsmdt

•

158

生产力与工作流

Orchestrates multi-phase implementation plans with incremental task loading and mandatory checkpoint validation.

关于

The Agent Coordination skill is a specialized execution specialist designed to manage complex, multi-step technical projects within Claude Code. It prevents context overload by loading tasks incrementally into TodoWrite, one phase at a time, ensuring the AI maintains focus on the current objective. The skill excels at identifying parallel execution opportunities, managing transitions between project phases with user-required checkpoints, and providing structured handling for blocked states. By synchronizing progress between implementation tools and plan documentation, it provides a robust framework for reliable, transparent, and structured software development workflows.

主要功能

158 GitHub stars
Automated synchronization between real-time progress and PLAN.md files
Structured blocked-state handling with actionable recovery and retry options
Incremental phase loading to manage LLM cognitive load and context window
Parallel task orchestration for concurrent agentic execution
Mandatory checkpoint validation at phase boundaries for user sign-off

使用场景

Executing complex architectural refactors that require sequential validation
Orchestrating parallel development tasks during the initial build phase of a startup project
Managing large-scale feature migrations across multiple service layers

关于

主要功能

158 GitHub stars
Automated synchronization between real-time progress and PLAN.md files
Structured blocked-state handling with actionable recovery and retry options
Incremental phase loading to manage LLM cognitive load and context window
Parallel task orchestration for concurrent agentic execution
Mandatory checkpoint validation at phase boundaries for user sign-off

使用场景

Executing complex architectural refactors that require sequential validation
Orchestrating parallel development tasks during the initial build phase of a startup project
Managing large-scale feature migrations across multiple service layers