msg-rocket is a customized CLI assistant powered by GitHub Copilot CLI. It streamlines your git workflow by leveraging AI to generate commit messages, conduct code reviews, ensuring your branches stay up-to-date, and enforcing team coding standards.

• AI-Generated Commit Messages: Automatically analyzes staged changes and generates descriptive commit messages using GitHub Copilot.
• Code Reviews: Performs AI-driven code reviews on your staged changes, with optional focus on performance or security.
• Clean Reports: suggestions for code cleanup and refactoring improvements.
• Standards Compliance: Validates your code changes against your team's specific coding standards.
• Smart Syncing: Keeps your branches up-to-date with the remote main branch, handling stashes and rebases automatically.
• Easter Egg: Includes a fun Matrix-style digital rain effect.

• Node.js: Version 14 or higher (must support ES modules).
• Git: Installed and available in your PATH.
• GitHub Copilot CLI: You must have the copilot command installed and authenticated.

npm install -g msg-rocket

1. Clone the repository:

   git clone https://github.com/Gramli/msg-rocket.git
   cd msg-rocket

2. Link the package globally:

   npm link

Now you can run the tool using the msg-rocket command.

Run the tool from your terminal within a git repository:

msg-rocket <command> [flags]

• commit: Generates a commit message for staged changes.

  • --t <ticket>: Appends ticket references (e.g., #123).
  • --f: Force commit without interactive confirmation.

• review: Generates a code review for staged changes.

  • --perf: Focus review on performance optimizations.
  • --sec: Focus review on security implications.

• clean: Generates a report with cleanup and refactoring suggestions.

• standard: Checks staged changes against your configured coding standards file.

• uptodate: Syncs the current branch with the remote main branch.

  • --m <branch>: Specify the main branch name (default: master).

• help: Displays help information.

You can configure msg-rocket by creating a .msgrocketrc file in your project root or user home directory. The file should contain valid JSON.

{
  "msgTemplate": "/path/to/my/template.txt",
  "teamCodingStandards": "/Users/developer/docs/coding-standards.md",
  "DEBUG": "false"
}

Generate a commit message with ticket references:

msg-rocket commit --t "JIRA-123" --t "#456"

Review staged changes for security issues:

msg-rocket review --sec

Sync with a specific main branch:

msg-rocket uptodate --m main

Check against coding standards:

Ensure teamCodingStandards is set in your .msgrocketrc, then run:

msg-rocket standard

When using free Copilot models, I noticed unstable behavior when passing prompts via files (e.g. copilot -p @prompt.md), including misinterpreting the file path as a user question.

In practice, the CLI behaved more consistently when the Reasoning Effort was set to High:

/model gpt-5-mini high

The following items describe potential future enhancements based on real usage and performance considerations:

• Token-efficient prompts
  Further reduce prompt size and Copilot processing time by dynamically trimming diffs, collapsing trivial changes, and reusing shared prompt fragments across commands.

• Branch name generation Generate meaningful git branch names from staged changes with optional prefix/postfix.

• Command chaining
  Allow combining multiple operations in a single run (e.g. review → clean → commit) while reusing context to avoid repeated Copilot calls.

• Partial diff analysis
  Support analyzing selected files or hunks to improve performance on large repositories and provide more targeted feedback.

• Optional Git hook integration
  Enable lightweight pre-commit or pre-push checks using existing commands, fully opt-in.

• Context reuse between commands
  Share analysis results across commands (e.g. commit + review) to minimize repeated reasoning and token usage.