What is the hybrid agent architecture?

The skill uses three specialized agents: pm-planner for strategy, pm-executor for task execution using the ReAct pattern, and pm-reflector for retrospective learning and estimation accuracy.

What database does this skill use for tracking?

It uses an SQLite database with an event-sourcing model, ensuring every change is tracked as an event for perfect audit trails and state reconstruction.

Can I sync my local tasks with GitHub Issues?

Yes, it supports bi-directional sync via pm_sync_pull and pm_sync_push, allowing you to import GitHub issues as tasks or push local updates to GitHub.

How does the PM skill integrate with Git?

It uses a LEVEL_1 Git-First design, enforcing branch naming conventions like {seq}-{type}-{description} and using magic words in commits (e.g., 'fixes #42') to automatically transition task statuses.

How does it handle long-term project context without hitting token limits?

The skill employs hierarchical summaries (L0-L3) and a '70% rule' that compresses context once it reaches 70% capacity, maintaining a 40-50% workspace buffer.

AI Project Management (PM)

Name: AI Project Management (PM)
Author: young1ll

byyoung1ll

Collaboration Tools

Streamlines project management workflows within Claude Code using MCP integration and hybrid AI agents.

About

This skill transforms Claude Code into a sophisticated project management hub by integrating MCP servers and event-sourcing via SQLite. It utilizes a hybrid agent architecture—Plan-and-Execute for strategy, ReAct for execution, and Reflexion for self-improvement—to manage tasks, sprints, and Git workflows seamlessly. By enforcing Git-first principles and offering deep GitHub synchronization, it ensures that project statuses, issues, and code changes remain perfectly aligned while optimizing token usage for efficient long-term collaboration.

Key Features

Event-sourcing based task and sprint management powered by SQLite and MCP.
Bi-directional synchronization between GitHub Issues and local project backlogs.
Hybrid AI Agent architecture (Planner, Executor, Reflector) for strategic project handling.
0 GitHub stars
Token-efficient context management using hierarchical summaries and the '70% Rule' for compression.
Standardized Git integration with automated branch naming and commit 'magic words' for status updates.

Use Cases

Analyzing developer velocity and project health through automated burndown data and Git hotspot analysis.
Automating sprint planning, daily standups, and retrospectives directly within the development terminal.
Maintaining a single source of truth between local task states and GitHub repository issues.

About

Key Features

Event-sourcing based task and sprint management powered by SQLite and MCP.
Bi-directional synchronization between GitHub Issues and local project backlogs.
Hybrid AI Agent architecture (Planner, Executor, Reflector) for strategic project handling.
0 GitHub stars
Token-efficient context management using hierarchical summaries and the '70% Rule' for compression.
Standardized Git integration with automated branch naming and commit 'magic words' for status updates.

Use Cases

Analyzing developer velocity and project health through automated burndown data and Git hotspot analysis.
Automating sprint planning, daily standups, and retrospectives directly within the development terminal.
Maintaining a single source of truth between local task states and GitHub repository issues.