Trekker CLI
A small issue tracker for AI agents: tasks, epics and dependencies live in a project SQLite file, with a local board.
- 88 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- Creates and edits a local SQLite file in the project
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/trekkerDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- Creates and edits a local SQLite file in the project
This is third-party code. Review the repository files before installing.
What it does
Trekker keeps tasks, epics, subtasks, dependencies and comments in a SQLite file inside the project, with no hosted service and no account. An agent pulls the task list, sees what is blocked and leaves notes for the next session. `trekker serve` opens a local board with todo, in progress, completed and won't fix columns. Main commands: `trekker ready` lists unblocked tasks and what each one frees, `trekker search` searches everything and `trekker dep add TREK-2 TREK-1` links tasks. Plugins for Claude Code and Codex teach the agent to search before starting, keep one task in progress and write a summary before closing. Other agents only need to be told to use Trekker: `trekker quickstart` prints a guide written for agents.
Who it is for. Developers whose agents lose context between sessions and forget what is done and what is waiting.
Good fit when
- An agent builds a large feature in pieces and must remember task order and dependencies.
- A new session starts with the question of what can be done right now, and `trekker ready` answers it.
- Epic progress should be visible on a local browser board.
- A file inside the project repository is wanted instead of a cloud tracker, with no account or hosting.
Not a fit when
- The team tracks work in a shared cloud tracker: Trekker is local and does not sync between people.
- Bun cannot be installed: the tool uses `bun:sqlite` and does not work without it.
- The agent needs an MCP server: Trekker has none and agents run the CLI directly.
Example request
Create an authentication epic with users table and login tasks, link them by dependency and show what is ready to work on.Limitations
Trekker runs on Bun, which must be installed first, even when installing through npm. Install with `bun install -g @obsfx/trekker` or `npm install -g @obsfx/trekker`. Data sits in a local project SQLite file and does not sync between machines. After installing, the Codex plugin requires trusting its hook with `/hooks`. The separate `trekker-claude-code` and `trekker-codex` plugin repositories are archived, and current plugins live in the main repository. The author notes the tool was built with AI help for personal use.
How to disable. Remove the package with `bun remove -g @obsfx/trekker` or `npm uninstall -g @obsfx/trekker` and disable the plugin. The project SQLite file is deleted by hand, and `trekker wipe` erases data.
FAQ
Is Bun needed when installing through npm?
Yes. Trekker uses `bun:sqlite`, so Bun is installed first and only then the package through `npm` or `bun`.
How do I connect it to Claude Code?
With `claude plugin marketplace add obsfx/trekker` and `claude plugin install trekker@trekker`. The plugin runs the CLI directly and there is no MCP server.
How do I connect it to Codex?
With `codex plugin marketplace add obsfx/trekker` and `codex plugin add trekker@trekker`, then trust the plugin hook through `/hooks`.
How do I teach another agent?
Tell it to track work in Trekker and run `trekker quickstart`. The command prints a full guide for agents.
Where can a person view the tasks?
`trekker serve` opens a local board with columns by status. No cloud service is needed.
Discussion
No questions or reviews yet.
Related
CLI toolsPlanDB is an issue tracker for AI agents: a dependency graph, atomic task claiming, notes and critical path analysis in SQLite.
A SQLite issue tracker for AI agents: sessions with handoff notes, dependencies, templates and Claude Code hooks
A task tracker for agents: tasks with context and result live in JSONL beside the code and outlive sessions
Yandex Tracker CLI
tracker
The tracker command-line tool for Yandex Tracker with JSON output and a ready SKILL.md workflow for agents
CLI and /trajex skill: Claude Code, Codex and Pi sessions in one SQLite index that an agent queries with JS
A Yandex Tracker CLI built for LLM agents: issues, comments, worklogs, checklists and bulk operations, with an agent skill
README in short
Covers why it was built, installation through Bun or npm, a quick start, plugin setup for Claude Code and Codex and the command list.