linear-cli

CLI for Linear — designed for AI Agents.

A command-line interface for the Linear project management tool that serves as a drop-in replacement for Linear's MCP (Model Context Protocol). Built primarily for AI Agents, it provides structured JSON output, meaningful exit codes, and zero interactive prompts.

Why linear-cli?

	Linear MCP	linear-cli
Connection	Persistent (stdio/SSE)	Stateless per-command
Protocol overhead	Handshake + negotiation	None
Framework support	MCP-compatible only	Any (shell exec)
Debugging	Complex	Run command in terminal
Composability	Limited	Pipes, xargs, scripts
Error handling	Protocol-level	Exit codes + JSON

Installation

Homebrew (macOS / Linux)

brew install k0walski/tap/linear-cli

Go install

go install github.com/enolalab/linear-cli@latest

Binary download

Download pre-built binaries from GitHub Releases.

Quick Start

# 1. Set your API key
export LINEAR_API_KEY=lin_api_xxxxx

# 2. Verify authentication
linear-cli auth whoami

# 3. List your teams
linear-cli team list

# 4. List issues
linear-cli issue list --team ENG --limit 5

# 5. Get issue details
linear-cli issue get ENG-123

Authentication

Set the LINEAR_API_KEY environment variable (recommended for AI Agents):

export LINEAR_API_KEY=lin_api_xxxxx

Or save it to the config file:

linear-cli auth login --token lin_api_xxxxx

You can get a personal API key from Linear Settings → API.

Priority: --api-key flag > LINEAR_API_KEY env var > config file (~/.config/linear-cli/config.yaml).

Output Format

All commands output a consistent JSON envelope:

Success:

{
  "success": true,
  "data": { ... },
  "pagination": {
    "hasNextPage": true,
    "endCursor": "cursor-string"
  }
}

Error:

{
  "success": false,
  "error": {
    "code": "NOT_FOUND",
    "message": "issue not found: ENG-999"
  }
}

Use --pretty for indented output when debugging.

Exit Codes

Code	Meaning	When
0	Success	Operation completed
1	General error	Unspecified error
2	Usage error	Invalid arguments or flags
3	Auth error	Missing or invalid API key
4	Not found	Resource doesn't exist
5	Rate limited	API quota exceeded
6	Network error	Connection failure
7	Permission denied	Insufficient access

Command Reference

Authentication & Config

Command	Description
auth login --token <key>	Save API key to config
auth whoami	Show current user
config set <key> <value>	Set config value
config get <key>	Get config value
config list	List all config (API key redacted)

Issues

Command	Key Flags
issue list	--team, --assignee, --status, --priority, --label, --limit, --cursor
issue get <ID>	Identifier like ENG-123
issue create	--title (req), --team, --description, --description-file, --priority, --assignee, --label, --estimate, --cycle, --project
issue update <ID>	--title, --status, --priority, --assignee, --label, --description, --description-file
issue search <query>	--limit

Comments

Command	Key Flags
comment list	--issue (required)
comment create	--issue (req), --body or --body-file

Teams & Users

Command	Key Flags
team list	--limit, --cursor
team get <key>	Team key like ENG
user list	--limit, --cursor
user get <id>	User UUID
user me	Alias for auth whoami

Workflow & Labels

Command	Key Flags
status list	--team (required)
status get <id>	State UUID
label list	--team
label create	--name (req), --team, --color

Projects & Cycles

Command	Key Flags
project list	--team, --limit, --cursor
project get <id>	Project UUID
project create	--name (req), --team, --description, --description-file
project update <id>	--name, --description, --state
cycle list	--team (required), --limit, --cursor
cycle get <id>	Includes issues

Documents

Command	Key Flags
doc list	--limit, --cursor
doc get <id>	Document UUID
doc search <query>	--limit

Attachments

Command	Key Flags
attachment upload	--issue (req), --file (req), --title

AI Agent Integration

As Shell Tools

AI Agents can invoke linear-cli commands as shell tools. Each command is:

• Stateless — no persistent connection needed
• Atomic — one command = one operation
• Predictable — JSON output with typed exit codes

Example: Agent Workflow

User: "Create 3 tasks for the current sprint in Engineering"

Agent executes:
1. linear-cli team list                          → find team ID
2. linear-cli cycle list --team ENG              → find current cycle
3. linear-cli status list --team ENG             → list workflow states
4. linear-cli issue create --title "Task 1" --team ENG
5. linear-cli issue create --title "Task 2" --team ENG
6. linear-cli issue create --title "Task 3" --team ENG

Example: Filtering & Search

# Issues assigned to me that are in progress
linear-cli issue list --assignee me --status "In Progress"

# Search for bugs
linear-cli issue search "login error" --limit 10

# Get all workflow states to know valid status values
linear-cli status list --team ENG

Pagination

For large result sets, use cursor-based pagination:

# First page
linear-cli issue list --team ENG --limit 10
# Response includes: "pagination": { "hasNextPage": true, "endCursor": "abc123" }

# Next page
linear-cli issue list --team ENG --limit 10 --cursor "abc123"

Default Team

Set a default team to avoid passing --team every time:

linear-cli config set default_team ENG
linear-cli issue list  # automatically uses team ENG

Configuration

Config file location: ~/.config/linear-cli/config.yaml

Key	Description	Example
api_key	Linear API key	lin_api_xxxxx
default_team	Default team key	ENG

Development

# Build
make build

# Run tests
make test

# Install locally
make install

# Clean
make clean

License

MIT