How do I see what happened in a previous session?

You can use the `trekker history` command to view a complete audit trail of changes, or `trekker search` to find specific context from previous task comments.

Can I use Trekker for complex projects with dependencies?

Yes, Trekker supports epics and subtasks, and includes a dependency system (`trekker dep add`) to ensure blockers are resolved before tasks are started.

Is Trekker safe for concurrent use?

Trekker uses SQLite with Write-Ahead Logging (WAL) to ensure multi-instance safety, allowing multiple agent instances to interact with the database concurrently.

Why should I use the --toon flag?

The --toon flag provides a programmatic, minimized output format that is highly token-efficient, saving cost and context space when listing many tasks.

How does Trekker differ from the built-in TodoWrite skill?

TodoWrite is conversation-scoped and loses state when context resets. Trekker uses a persistent SQLite database that survives across multiple sessions and terminal restarts.

Trekker

Name: Trekker
Author: obsfx

byobsfx

0•

生产力与工作流

Manages persistent task memory and issue tracking for AI agents across multiple sessions using a SQLite database.

Trekker is a robust issue-tracking system designed specifically for AI agents, providing a persistent memory layer for complex, multi-session development projects. Unlike conversation-scoped tools, Trekker stores tasks, epics, and subtasks in a local SQLite database, allowing agents to recover context, track progress, and manage dependencies even after a terminal reset or session timeout. By maintaining a detailed audit log and full-text search capabilities, it ensures that project history is never lost, preventing duplicate efforts and facilitating seamless handoffs between AI interactions.

主要功能

01Full-text search for recalling past decisions, code changes, and investigation notes

02Detailed history and audit trails to track changes across sessions

030 GitHub stars

04Hierarchical task management including Epics, Tasks, and Subtasks

05Persistent SQLite storage for multi-session task and context retention

06Token-efficient programmatic output mode via the --toon flag

使用场景

01Managing project milestones and feature dependencies for long-term coding tasks

02Documenting bug investigations and architectural decisions for future AI agents

03Resuming complex development workflows across multiple AI context resets

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add obsfx/trekker-claude-code trekker

For use in Claude.ai and ChatGPT