How does this skill handle context window limits?

It utilizes a handoff coordination system that saves the current state to a JSON file and triggers a fresh session once the context usage reaches a defined threshold.

Where is the session state stored?

Session state and progress are persisted in local files like .claude-session-state.json and the feature_list.json within your project directory.

What is the difference between INIT and CONTINUE sessions?

An INIT session scaffolds a new project and generates a feature list, while a CONTINUE session resumes work on existing features using saved handoff data.

Does it support Test-Driven Development (TDD)?

Yes, the continuation logic is specifically built to follow TDD patterns, selecting one feature at a time to write failing tests, implement code, and refactor.

Autonomous Session Manager

Name: Autonomous Session Manager
Author: adaptationio

byadaptationio

生産性とワークフロー

Manages the complete lifecycle of autonomous coding sessions, including state preservation, session detection, and seamless handoffs.

概要

The Autonomous Session Manager is a foundation-layer skill designed to orchestrate complex, long-running development tasks by managing session transitions for AI agents. It intelligently detects whether a project requires initialization, implementation, or a resume-from-handoff operation, ensuring that context is never lost between runs. By implementing logic to select the next feature and handle graceful shutdowns, it allows autonomous agents to work reliably across multiple sessions until all project requirements are verified and complete.

主な機能

Standardized TDD-driven feature implementation logic
Automatic session type detection (INIT, CONTINUE, or COMPLETE)
Configurable auto-continuation for multi-session workflows
State preservation through JSON-based handoff packages
Graceful shutdown with comprehensive project status reporting
0 GitHub stars

ユースケース

Building large-scale applications that exceed single-session context windows
Automating the initialization of new repositories with predefined feature sets
Ensuring consistent progress tracking across independent autonomous coding runs

概要

主な機能

Standardized TDD-driven feature implementation logic
Automatic session type detection (INIT, CONTINUE, or COMPLETE)
Configurable auto-continuation for multi-session workflows
State preservation through JSON-based handoff packages
Graceful shutdown with comprehensive project status reporting
0 GitHub stars

ユースケース

Building large-scale applications that exceed single-session context windows
Automating the initialization of new repositories with predefined feature sets
Ensuring consistent progress tracking across independent autonomous coding runs