session-manager

Analyze and manage local Claude Code session files in JSONL format.

1|1|Updated Jul 7, 2026
One-click install
npx skills add https://github.com/hicucu/my-poor-ai --skill session-manager-hicucu
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: session-manager
Source: https://github.com/hicucu/my-poor-ai/tree/main/skills/session-manager
Command: npx skills add https://github.com/hicucu/my-poor-ai --skill session-manager-hicucu

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill solves the clutter and management overhead of local Claude Code session files by providing a centralized interface to list, rename, and safely delete session history.

Core Features & Use Cases

  • Parallel Session Analysis: Uses up to 10 subagents to concurrently parse and summarize session metadata from JSONL files.
  • Intelligent Cleanup: Supports bulk deletion via natural language conditions (e.g., date-based or project-based) with mandatory safety confirmation.
  • Use Case: Quickly identify and remove stale development sessions from specific projects to keep your workspace organized and reduce noise in your session history.

Quick Start

Use the session-manager skill to list all your recent Claude Code sessions and identify those that are no longer needed.

Frequently Asked Questions about session-manager

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I clean up Claude Code session files and remove stale history?

To clean up Claude Code session files, you can use parallel analysis to parse JSONL session logs and execute conditional batch deletion. This allows you to filter sessions by project or date and safely remove stale development history.

Can I rename specific Claude Code sessions from the command line?

Yes, you can rename Claude Code sessions by utilizing administrative commands that operate directly on local JSONL logs. This provides a centralized interface to update session metadata and keep your workspace organized.

How does parallel analysis work when parsing JSONL session logs?

Parallel analysis works by deploying up to 10 subagents to concurrently extract and summarize metadata from JSONL session logs. This mechanism significantly speeds up the identification of sessions across different projects.

Do I need Python to manage and traverse local Claude Code session logs?

Yes, Python is required for file system traversal and data extraction, while Bash is used for process orchestration. These dependencies ensure efficient parallel parsing and administrative cleanup of your session history.

What is the best way to batch delete Claude Code sessions based on project names?

The best way to batch delete sessions is by using natural language conditions for project-specific filtering, which includes a mandatory safety confirmation before removing the JSONL files. This prevents accidental data loss during cleanup.

Are there limitations when filtering Claude Code session history for cleanup?

A primary limitation during session cleanup is the mandatory safety confirmation required before executing any batch deletion. Users must also ensure their Python and Bash environments are correctly configured for JSONL file traversal.