How does the batching system work in this skill?

The skill executes a default of three tasks from your plan at a time. After each batch, it stops to report progress and verify results, waiting for human feedback before moving to the next set of tasks.

Does this skill help with finishing a development branch?

Yes, once all tasks in the plan are complete, the skill automatically triggers a requirement to use the 'finishing-a-development-branch' skill to verify the final output.

When is it best to use the executing-plans skill?

Use this skill whenever you have a detailed, written implementation plan that needs to be executed precisely, especially for complex tasks where you want to review progress at regular intervals.

What happens if Claude encounters an error during execution?

The skill is programmed to stop immediately if it hits a blocker, such as a failed test or an unclear instruction, to prevent guesswork and ensure the plan remains on track.

Agentic Plan Execution

Name: Agentic Plan Execution
Author: noeljackson

bynoeljackson

0•

生产力与工作流

Executes complex implementation plans through a structured, batch-based workflow with mandatory human review checkpoints.

This skill provides a disciplined framework for Claude to execute pre-written implementation plans without losing context or deviating from architectural goals. By processing tasks in batches—typically three at a time—the skill ensures that progress is validated at every stage, allowing for human intervention and feedback before proceeding. It emphasizes a 'stop-and-ask' philosophy for blockers, making it an essential tool for high-stakes development tasks where precision and systematic verification are required.

主要功能

01Pre-execution critical review to identify plan gaps or risks

02Batch-based task execution to maintain focus and quality

030 GitHub stars

04Structured reporting for efficient feedback loops

05Integrated verification steps for every completed task

06Automatic hand-off to branch finalization workflows

使用场景

01Managing long-running development tasks that require human-in-the-loop checkpoints

02Executing multi-file architectural refactors from a written plan

03Implementing complex feature sets with specific verification requirements

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add noeljackson/agentic executing-plans

For use in Claude.ai and ChatGPT

主要功能

01Pre-execution critical review to identify plan gaps or risks

02Batch-based task execution to maintain focus and quality

030 GitHub stars

04Structured reporting for efficient feedback loops

05Integrated verification steps for every completed task

06Automatic hand-off to branch finalization workflows

使用场景

01Managing long-running development tasks that require human-in-the-loop checkpoints

02Executing multi-file architectural refactors from a written plan

03Implementing complex feature sets with specific verification requirements

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add noeljackson/agentic executing-plans

For use in Claude.ai and ChatGPT