# Agent Skills
`linear-cli` includes Agent Skills for AI coding assistants. Skills provide contextual documentation that agents can load when performing Linear tasks.
## Installation
```bash
# Install all skills
npx skills add Finesssee/linear-cli
# Install specific skill
npx skills add Finesssee/linear-cli --skill linear-list
# Install globally (available in all projects)
npx skills add Finesssee/linear-cli -g
```
## Available Skills (27 total)
### Issues
| `linear-list` | List and get issues |
| `linear-create` | Create issues |
| `linear-update` | Update issues (status, priority, assignee, labels) |
| `linear-workflow` | Start/stop work, get current issue context |
### Git
| `linear-git` | Branches, checkout, context |
| `linear-pr` | Create GitHub PRs |
### Planning
| `linear-projects` | Manage projects |
| `linear-roadmaps` | View roadmaps |
| `linear-initiatives` | High-level tracking |
| `linear-cycles` | Sprint cycles |
### Organization
| `linear-teams` | Teams and users |
| `linear-labels` | Label management |
| `linear-relations` | Issue relationships (blocks, parent/child) |
| `linear-templates` | Issue templates |
### Operations
| `linear-bulk` | Bulk operations |
| `linear-export` | Export to CSV/Markdown |
| `linear-triage` | Triage inbox |
| `linear-favorites` | Quick access |
### Tracking
| `linear-metrics` | Velocity, burndown, progress |
| `linear-history` | Issue activity logs |
| `linear-time` | Time tracking |
| `linear-watch` | Watch for updates |
### Other
| `linear-search` | Search issues and projects |
| `linear-notifications` | Manage notifications |
| `linear-documents` | Documentation |
| `linear-uploads` | Download attachments |
| `linear-config` | Auth, API keys, workspaces, diagnostics |
## Supported Agents
Skills work with any agent that supports the [Agent Skills](https://agentskills.io) format:
- Claude Code
- OpenAI Codex
- Cursor
- Amp
- Roo Code
- Gemini CLI
- And many more
## Why Skills?
Skills are 10-50x more token-efficient than MCP tools:
- **MCP tools**: Each API call returns full JSON, uses many tokens
- **Skills**: Agent learns commands once, uses CLI directly
## Viewing Installed Skills
```bash
# List installed skills
npx skills list
# List globally installed
npx skills list -g
```
## Skill Contents
Each skill contains:
- **Frontmatter**: Name, description, allowed tools
- **Commands**: CLI commands with examples
- **Flags**: Agent-optimized flags (`--output json`, `--compact`, etc.)
- **Exit codes**: For error handling
- **Workflows**: Common task patterns
Example skill structure:
```yaml
---
name: linear-list
description: List and get Linear issues...
allowed-tools: Bash
---
# List/Get Issues
\`\`\`bash
linear-cli i list --output json
\`\`\`
```
## Updating Skills
```bash
# Check for updates
npx skills check
# Update all skills
npx skills update
```
## Removing Skills
```bash
# Remove specific skill
npx skills remove --skill linear-list
# Remove all linear-cli skills
npx skills remove Finesssee/linear-cli
```