Linear CLI

An agent-native Linear runtime for the current v3 execution model, with stable JSON contracts, startup discovery, dry-run previews, timeout-aware write semantics, source-adjacent intake, and git/jj workflow integration.

Recommended Agent Loop

When using this CLI from an agent runtime, prefer this order:

1. Discover command traits with linear capabilities; use --compat v1 only when an older consumer still expects the trimmed legacy shape
2. Read Linear state with default-JSON core surfaces or --json
3. Preview writes with --dry-run --json when available
4. Apply writes on the default machine-readable surface, then inspect operation, receipt, and error.details
5. Inspect exit codes and error.details instead of parsing styled terminal text

Prompt-driven human/debug flows are secondary and explicit. When a command supports prompts or editor entry, pass --profile human-debug --interactive; otherwise missing required inputs fail fast.

Agent-safe execution semantics are now the default runtime behavior. --text and --profile human-debug are the explicit human/debug escape hatches for maintainers, and --profile agent-safe remains accepted for compatibility with older automation.

When upstream tooling hands the runtime a normalized Slack, ticket, or similar source envelope, prefer --context-file, add --apply-triage if that envelope already contains deterministic team/state/label hints, and choose --autonomy-policy explicitly when the wrapper needs suggest-only or preview-required staging.

Recommended supporting docs:

• ../../docs/agent-first.md
• ../../docs/v2-to-v3-migration-cookbook.md
• ../../docs/json-contracts.md
• ../../docs/stdin-policy.md

Prerequisites

The linear command must be available on PATH. To check:

linear --version

If not installed, follow the instructions at: https://github.com/kyaukyuai/linear-cli?tab=readme-ov-file#install

Best Practices for Markdown Content

When working with issue descriptions or comment bodies that contain markdown, prefer file-based flags for existing files and stdin for generated pipeline content:

• Use --description-file for issue create and issue update commands when the content already exists on disk
• Use --body-file for comment add and comment update commands when the content already exists on disk
• Pipe stdin for generated markdown, for example cat description.md | linear issue create --title "My Issue" --team ENG

Why avoid large inline flags:

• Ensures proper formatting in the Linear web UI
• Avoids shell escaping issues with newlines and special characters
• Prevents literal \n sequences from appearing in markdown
• Makes it easier to work with multi-line content in scripts and pipelines

Example workflow:

cat > /tmp/description.md <<'EOF'
## Summary

- First item
- Second item

## Details

This is a detailed description with proper formatting.
EOF

linear issue create --title "My Issue" --description-file /tmp/description.md

cat /tmp/description.md | linear issue create --title "My Issue" --team ENG

linear issue comment add ENG-123 --body-file /tmp/comment.md

Only use inline flags (--description, --body) for simple, single-line content.

Available Commands

linear auth               # Manage Linear authentication
linear issue              # Manage Linear issues
linear team               # Manage Linear teams
linear project            # Manage Linear projects
linear project-update     # Manage project status updates
linear cycle              # Manage Linear team cycles
linear milestone          # Manage Linear project milestones
linear initiative         # Manage Linear initiatives
linear initiative-update  # Manage initiative status updates (timeline posts)
linear label              # Manage Linear issue labels
linear document           # Manage Linear documents
linear notification       # Manage Linear notifications
linear webhook            # Manage Linear webhooks
linear workflow-state     # Manage Linear workflow states
linear user               # Manage Linear users
linear project-label      # Manage Linear project labels
linear config             # Interactively generate .linear.toml configuration
linear schema             # Print the GraphQL schema to stdout
linear api                # Make a raw GraphQL API request
linear capabilities       # Describe the agent-facing command surface
linear resolve            # Resolve references without mutating Linear

Reference Documentation

• auth - Manage Linear authentication
• issue - Manage Linear issues
• team - Manage Linear teams
• project - Manage Linear projects
• project-update - Manage project status updates
• cycle - Manage Linear team cycles
• milestone - Manage Linear project milestones
• initiative - Manage Linear initiatives
• initiative-update - Manage initiative status updates (timeline posts)
• label - Manage Linear issue labels
• document - Manage Linear documents
• notification - Manage Linear notifications
• webhook - Manage Linear webhooks
• workflow-state - Manage Linear workflow states
• user - Manage Linear users
• project-label - Manage Linear project labels
• config - Interactively generate .linear.toml configuration
• schema - Print the GraphQL schema to stdout
• api - Make a raw GraphQL API request
• capabilities - Describe the agent-facing command surface
• resolve - Resolve references without mutating Linear

For curated examples of organization features (initiatives, labels, projects, bulk operations), see organization-features.

Discovering Options

To see available subcommands and flags, run --help on any command:

linear --help
linear issue --help
linear issue list --help
linear issue create --help

Each command has detailed help output describing all available flags and options.

For machine-readable discovery, prefer:

linear capabilities
linear capabilities --compat v1

Using the Linear GraphQL API Directly

Prefer the CLI for all supported operations. The api command should only be used as a fallback for queries not covered by the CLI.

Check the schema for available types and fields

Write the schema to a tempfile, then search it:

linear schema -o "${TMPDIR:-/tmp}/linear-schema.graphql"
grep -i "cycle" "${TMPDIR:-/tmp}/linear-schema.graphql"
grep -A 30 "^type Issue " "${TMPDIR:-/tmp}/linear-schema.graphql"

Make a GraphQL request

Important: GraphQL queries containing non-null type markers (e.g. String followed by an exclamation mark) must be passed via heredoc stdin to avoid escaping issues. Simple queries without those markers can be passed inline.

linear api '{ viewer { id name email } }'

linear api --variable teamId=abc123 <<'GRAPHQL'
query($teamId: String!) { team(id: $teamId) { name } }
GRAPHQL

linear api --variable term=onboarding <<'GRAPHQL'
query($term: String!) { searchIssues(term: $term, first: 20) { nodes { identifier title state { name } } } }
GRAPHQL

linear api --variable first=5 <<'GRAPHQL'
query($first: Int!) { issues(first: $first) { nodes { title } } }
GRAPHQL

linear api --variables-json '{"filter": {"state": {"name": {"eq": "In Progress"}}}}' <<'GRAPHQL'
query($filter: IssueFilter!) { issues(filter: $filter) { nodes { title } } }
GRAPHQL

linear api '{ issues(first: 5) { nodes { identifier title } } }' | jq '.data.issues.nodes[].title'

Advanced: Using curl directly

For cases where you need full HTTP control, use linear auth token:

curl -s -X POST https://api.linear.app/graphql \
  -H "Content-Type: application/json" \
  -H "Authorization: $(linear auth token)" \
  -d '{"query": "{ viewer { id } }"}'

依赖说明

运行环境

• Agent平台: 支持SKILL.md的任意AI Agent(Claude Code / Cursor / Codex / Gemini CLI等)
• 操作系统: Windows / macOS / Linux

第三方依赖

API Key 配置

• 本Skill基于Markdown指令,无需额外API Key(除内容中明确标注的外部API)

可用性分类

• 分类: MD+EXEC(纯Markdown指令,部分功能需要exec命令行执行能力)
• 说明: 基于Markdown的AI Skill,通过自然语言指令驱动Agent执行任务