allyas

Personal shell aliases for macOS - Source Repository

This is the source code repository for allyas. The Homebrew formula is maintained separately in homebrew-allyas.

What is this?

allyas is a collection of shell aliases that can be installed and managed via Homebrew. This repository contains:

• allyas.sh - The shell aliases file
• .github/workflows/release.yml - Automation that updates the Homebrew tap

Installation (for users)

Don't install from this repo directly! Use the Homebrew tap instead:

brew tap rezkam/allyas
brew install allyas

For more details, see homebrew-allyas.

Usage

After sourcing allyas.sh (usually via the Homebrew shim), run the allyas command to print every alias/function along with the section headings they live under. This is the quickest way to discover the shortcuts that are available in your shell session.

LLM-Powered Commands

allyas includes LLM-powered security analysis tools that require one of the following CLI tools:

• codex - OpenAI Codex (default, fast ~6s)
• claude - Anthropic Claude (fastest ~5s with Haiku)
• gemini - Google Gemini (~29s, slower)

portswhy

Analyze all listening TCP ports on your system and identify suspicious processes:

portswhy

This command:

1. Scans all listening TCP ports
2. Identifies the processes using those ports
3. Checks code signatures for each executable
4. Uses LLM to analyze security implications and provide recommendations

LLM Configuration

Switch between LLM providers:

llm-use codex   # OpenAI Codex (default, fast)
llm-use claude  # Anthropic Claude (fastest with Haiku model)
llm-use gemini  # Google Gemini
llm-list        # Show available LLMs and current configuration

Model Customization

Override default models via environment variables:

# Temporary override for current session
export ALLYAS_CODEX_MODEL="gpt-4o"
export ALLYAS_CLAUDE_MODEL="sonnet"
export ALLYAS_GEMINI_MODEL="gemini-2.0-flash-exp"

# Then run your command
portswhy

To make changes permanent, add to your ~/.zshrc or ~/.bashrc:

# In ~/.zshrc
export ALLYAS_LLM="claude"              # Set default LLM
export ALLYAS_CLAUDE_MODEL="sonnet"     # Use Claude Sonnet instead of Haiku

Default models:

• Codex: gpt-5-codex
• Claude: haiku (optimized for speed)
• Gemini: gemini-2.5-flash

Performance note: Claude Haiku and OpenAI Codex are significantly faster (~5-6s) than Gemini (~29s) for typical analyses.

Features

• Automated releases: Tag once, everything else is automatic
• Clean separation: Source code separate from distribution
• No self-modifying repos: Tap repo only updated by external triggers
• Version controlled: All changes tracked in git
• Cross-shell compatible: Works with both bash and zsh
• Cross-architecture: Works on both Apple Silicon and Intel Macs