emdash-ai 0.1.176

Graph-based coding intelligence system - The 'Senior Engineer' Context Engine

emdash-ai

Graph-based coding intelligence system - The 'Senior Engineer' Context Engine

Quick Install (macOS)

brew install uv cmake
uv tool install emdash-ai

Quick Install (Linux)

# Install dependencies
sudo apt-get install cmake  # Debian/Ubuntu
# or: sudo dnf install cmake  # Fedora

# Install uv
curl -LsSf https://astral.sh/uv/install.sh | sh

# Add uv to PATH (add to ~/.bashrc or ~/.zshrc)
export PATH="$HOME/.cargo/bin:$PATH"

uv tool install emdash-ai

Note: After installation, you may need to reload your shell config:

source ~/.zshrc  # or ~/.bashrc

Update

uv tool upgrade emdash-ai

Requirements

• Python 3.10+ (3.10-3.13 recommended)
• cmake (for building the graph database)

Usage

em              # Start the AI coding agent
emdash index    # Index your codebase
emdash agent    # Start the AI agent (alternative)

Slash Commands

Once inside the agent REPL, use these slash commands:

Mode Switching

Command	Description
/plan	Switch to plan mode (explore codebase, create plans)
/code	Switch to code mode (execute file changes)
/mode	Show current operating mode

Generation & Research

Command	Description
/projectmd	Generate PROJECT.md describing your codebase architecture
/pr [url]	Review a pull request (e.g., /pr 123 or /pr https://github.com/org/repo/pull/123)
/research [goal]	Deep research on a topic (e.g., /research How does auth work?)

Status & Context

Command	Description
/status	Show index and PROJECT.md status
/context	Show current context frame (tokens, reranked items)
/compact	Compact message history using LLM summarization
/spec	Show current specification from plan mode

Todo Management

Command	Description
/todos	Show current agent todo list
/todo-add [title]	Add a todo item for the agent

Session Management

Command	Description
/session	Interactive session menu
/session save [name]	Save current session
/session load [name]	Load a saved session
/session list	List all saved sessions
/session delete [name]	Delete a session
/reset	Reset session state

Configuration

Command	Description
/agents	Manage agents (create, show, edit, delete)
/rules	Manage rules that guide agent behavior
/skills	Manage reusable skills
/hooks	Manage event hooks (triggers on tool_start, session_end, etc.)
/mcp	Manage MCP servers
/registry	Browse and install community plugins, skills, rules, agents, and verifiers
/setup	Interactive setup wizard

Verification

Command	Description
/verify	Run verification checks on current work
/verify-loop [task]	Run task in loop until all verifications pass

Authentication & Diagnostics

Command	Description
/auth	GitHub authentication (login, logout, status)
/doctor	Check environment and diagnose issues

Help & Exit

Command	Description
/help	Show available commands
/quit	Exit the agent (also /exit, /q)

Keyboard Shortcuts

Shortcut	Action
Ctrl+Space	Auto-complete commands
Alt+Enter	Multi-line input
Ctrl+V	Paste image from clipboard
@filename	Reference a file in your prompt

Features

• Graph-based code analysis
• Semantic code search
• AI-powered code exploration
• MCP server integration
• Session persistence
• Custom agents, rules, and skills
• Event hooks for automation
• Verification loops
• Plugin system for bundled workflows

Configuration

Configuration files are stored in .emdash/ in your project root:

.emdash/
├── agents/       # Custom agent definitions
├── rules/        # Coding rules and guidelines
├── skills/       # Reusable skills
├── plugins/      # Installed plugin manifests
├── sessions/     # Saved sessions
├── hooks.json    # Event hooks configuration
├── plugins.json  # Installed plugins tracking
└── verifiers.json # Verification configurations

Global config at ~/.config/emdash/:

~/.config/emdash/
├── mcp.json      # MCP server configuration
└── config.json   # Global settings

Troubleshooting

Build fails with "cmake: command not found"

If you see Failed to build kuzu with cmake errors:

# macOS
brew install cmake

# Linux (Debian/Ubuntu)
sudo apt-get install cmake

# Then retry
uv tool install emdash-ai

"permission denied: em"

If you get zsh: permission denied: em or similar:

# Check if uv tool installed it
uv tool list

# Reinstall
uv tool uninstall emdash-ai && uv tool install emdash-ai

Command not found

If em or emdash is not found, add ~/.local/bin to your PATH:

# For zsh (add to ~/.zshrc)
export PATH="$HOME/.local/bin:$PATH"

# For bash (add to ~/.bashrc or ~/.bash_profile)
export PATH="$HOME/.local/bin:$PATH"

# Then reload
source ~/.zshrc  # or ~/.bashrc

Debug installation

# Check which em is being found
which em
type em

# Check permissions
ls -la ~/.local/bin/em

# Check installed version
uv tool list

Uninstall

uv tool uninstall emdash-ai

Running Terminal-Bench

Run the terminal-bench benchmark to evaluate agent performance.

Prerequisites

• Docker running (Colima or Rancher Desktop)
• Harbor installed: pip install harbor

Setup

cd benchmarks
uv venv .venv
uv pip install -r requirements.txt

Run

./benchmarks/scripts/run.sh

With custom settings:

MODEL="qwen3-vl-235b" PARALLEL=2 ./benchmarks/scripts/run.sh

Variable	Default	Description
MODEL	claude-haiku-4-5	Model to benchmark
PARALLEL	4	Concurrent tasks
DATASET	terminal-bench@2.0	Dataset to run