Download the PHP package idoko-emmanuel/ticket-context-cli without Composer

ticket-context-cli

A personal CLI tool that pulls your active Jira sprint tickets and generates rich Markdown context files optimised for pasting into Claude Code (or any AI coding assistant).

Run it from any project repo on your machine. It automatically scopes to the right Jira project and git branch.

Why this tool?

Switching to a new ticket used to mean 5 browser tabs open, copy-pasting ticket details into Claude, manually creating a branch, and hoping I hadn't missed any acceptance criteria.

Now one command (tix branch LTN-42) creates the branch, fetches the full ticket, and saves a structured Markdown context file. Claude Code reads it automatically at session start via its skills system — so the AI already knows exactly what I'm building before I type a word.

How it works

Requirements

• PHP 8.2+ — php.net/downloads or via your package manager
• Composer — getcomposer.org/download (the install script handles this automatically)
• Git

Installation

One-line install

The install script checks for PHP, installs Composer if it's missing, clones the repo, sets everything up, and adds the tix shell function automatically.

Then reload your shell:

Alternative: Composer global install

If you already have Composer installed and prefer managing tools through it:

Ensure ~/.composer/vendor/bin is in your PATH after installing.

Manual install

If you prefer to run each step yourself:

1. Install Composer (skip if already installed)

2. Clone and install

3. Add the global shell function

Add this to your ~/.zshrc or ~/.bashrc:

Then reload:

You can now run tix <command> from any directory on your machine.

Uninstallation

This will:

• Remove ~/.local/share/ticket-context-cli
• Remove the tix shell function from ~/.zshrc / ~/.bashrc
• Prompt whether to also delete your saved Jira credentials (~/.config/ticket-context/)

Setup

Configure your Jira credentials

You will be prompted for:

Credentials are saved to ~/.config/ticket-context/config.json with chmod 600 — never inside any project repo.

Verify the connection

Shows your credentials status, live connection test, and current project link.

Per-project setup

Run once from the root of each project repo you want to track:

This creates .ticket-context.json in the repo root and adds .claude/skills/ticket-context/*-context.md to .gitignore.

Your project key is the prefix of every ticket — e.g. LTN from LTN-42. You can also find it in your Jira board URL:

Daily workflow

1. See your active sprint tickets

Shows a table of tickets assigned to you in the current sprint, filtered to the linked project. Use --all to see across all projects, --json for raw output.

2. Start a ticket — create a branch and save context

This will:

• Create git branch feature/ltn-42-{kebab-slug-of-summary}
• Link the ticket to that branch in .ticket-context.json
• Fetch the full ticket from Jira and save it to .claude/skills/ticket-context/LTN-42-context.md
• Create .claude/skills/ticket-context/SKILL.md so Claude loads the context automatically

Also transition the ticket when branching

Link multiple tickets to one branch

All tickets are linked to the branch. A combined context file is saved with each ticket as a separate section.

Skip branch creation (just save context)

3. Generate or refresh context

With no arguments, auto-detects tickets from the current git branch, saves the context to .claude/skills/ticket-context/ and prints a confirmation. Explicit keys work the same way:

Optionally transition at the same time:

4. Move a ticket to a different status

Auto-detects tickets from the current branch and moves all of them. You can also pass explicit keys — no quotes needed for multi-word statuses:

5. Check available statuses

Lists all statuses available in the linked project — useful when you're not sure of the exact name to pass to --transition or tix move.

Claude Code integration

A Claude skill file is automatically written to .claude/skills/ticket-context/SKILL.md the first time any tix command is run in a linked project. Claude Code reads skills from .claude/skills/ at session start, so ticket context is loaded automatically — no manual configuration needed.

Manual use

You can also reference a context file explicitly at any time:

Command reference

File reference

Contributing

1. Create a feature branch from main:

2. Make your changes and commit them.

3. Push your branch and open a pull request against main:

4. Ensure the CI checks (lint, unit tests, feature tests) pass before requesting a review.

License

MIT