slash-agent brings a zero-overhead LLM copilot to your Bash shell
slash-agent is an ultra-lightweight, on-demand AI coding assistant that integrates natively into an active Bash shell with no background daemons or idle memory usage, supporting both local Ollama models and cloud providers like OpenAI and Azure OpenAI.
Score breakdown
slash-agent removes the need for a persistent background process to get LLM assistance in the terminal, making AI coding help available on-demand with zero idle resource cost and full support for local private models.
- 01Triggered by typing `/agent` in the terminal — no background process runs when idle
- 02Captures recent terminal context via tmux pane scrollback or command history to diagnose errors and execute commands
- 03Automatically syncs `cd` directory changes and environment exports back to the parent shell on exit
slash-agent, published on GitHub by akatzmann, is an ultra-lightweight AI coding assistant designed to live natively inside an active Bash shell. Unlike persistent AI tools, it runs no background daemons and consumes no idle memory — it activates only when the user types `/agent`. At that point it immediately captures recent terminal context from tmux pane scrollback or command history to help diagnose errors, edit files, and run commands.
A notable design detail is that directory changes (`cd`) and environment exports made during an agent session are automatically synced back to the parent shell when the agent exits, preserving shell state seamlessly.
A notable design detail is that directory changes (`cd`) and environment exports made during an agent session are automatically synced back to the parent shell when the agent exits, preserving shell state seamlessly. The tool is LLM-agnostic: it supports fully local, private inference via Ollama (requiring no API keys and keeping all data on-device) as well as cloud providers OpenAI and Azure OpenAI. Installation is a single `curl` command that clones the repo to `~/.slash-agent`, configures a Python virtual environment, installs requirements, and registers shell integration in the appropriate profile file (`~/.zshrc`, `~/.bashrc`, `~/.bash_profile`, or `config.fish`).
Key facts
- 01Triggered by typing `/agent` in the terminal — no background process runs when idle
- 02Captures recent terminal context via tmux pane scrollback or command history to diagnose errors and execute commands
- 03Automatically syncs `cd` directory changes and environment exports back to the parent shell on exit
- 04Supports local offline models via Ollama (no API keys, no data leaves the machine) and cloud providers OpenAI and Azure OpenAI
- 05One-line `curl` installer supports Bash, Zsh, Ksh, and Fish shells
- 06Clones repo to `~/.slash-agent` and configures a Python virtual environment automatically
- 07Published on GitHub by akatzmann with 6 stars at time of posting
Topics
Summary and scoring are generated automatically from the original article. We always link back to the publisher and never republish images or paywalled content. Last processed Jun 20, 2026 · 08:55 UTC. How this works →