When should I create a sprint vs. working directly?

The skill recommends creating a sprint in prd.json for 5 or more related tasks. For fewer tasks or creative design work, it suggests working directly to avoid unnecessary ceremony.

What is the Core Task & Sprint Manager?

It is a specialized capability for Claude Code that manages development tasks using a hybrid system of native commands and a persistent prd.json file for long-term tracking.

How does it save context tokens?

Instead of reading the entire project history, it uses Grep and targeted Read commands with offsets to find specific stories, keeping Claude's context window focused on the current task.

What is resolution learning?

Resolution learning is a feature that prompts Claude to document how a specific bug was fixed using a [PATTERN]: [FIX] format, creating a searchable knowledge base of technical solutions.

How does it handle very large task lists?

The skill includes an archive trigger that automatically suggests moving completed stories to a separate archive file when the prd.json exceeds 50KB or 500 lines.

Core Task & Sprint Manager

Name: Core Task & Sprint Manager
Author: djnsty23

bydjnsty23

•

Productivity & Workflow

Manages project workflows using a hybrid system of native Claude tasks and persistent JSON-based sprint tracking.

The Core Task & Sprint Manager provides a robust framework for autonomous development within Claude Code, bridging the gap between transient session tasks and long-term project documentation. It implements a structured prd.json schema to track sprints, stories, and bug resolutions while optimizing context usage through targeted file reading and grepping. By automating the transition between active work and git-tracked history, it ensures that project progress is documented without sacrificing the speed of native AI task execution, making it ideal for teams needing structured tracking without manual overhead.

Key Features

01Context-optimized file access using Grep and targeted reads to save tokens

02Automated sprint tracking with status reporting and progress summaries

03Standardized resolution learning pattern to document technical fixes and prevent regressions

04Hybrid task management combining native session tasks with persistent prd.json storage

05Automatic archive management for large project files to maintain peak performance

062 GitHub stars

Use Cases

01Managing complex development sprints with multiple interdependent stories and task loops

02Maintaining a lean context window by offloading task history to structured archives

03Documenting bug fix patterns and architectural decisions in a git-tracked history

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add djnsty23/claude-auto-dev core

For use in Claude.ai and ChatGPT

Download Skill