Linear Claude Agent — Development skill for Claude Code
Assign a Linear issue to Claude and get a real, steerable Claude Code session on your own machine.
How to install Linear Claude Agent
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open julian-pineiro/linear-claude-agent and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Linear Claude Agent does
Assign a Linear issue to Claude and get a real, steerable Claude Code session on your own machine.
Alternatives in Development
- Impact Linear — Analyze Linear issue impact and dependencies 512 ★
- Alethe Agents — A local-first desktop workspace for running, organizing, and resuming multiple coding agents and shells with r 511 ★
- MCP Server Linear — Linear issue tracker integration 347 ★
README
linear-claude-agent
Assign a Linear issue to Claude and a real Claude Code session starts on your own machine — one you can open and steer from Claude Desktop, claude.ai/code or your phone.
Claude Code has no native Linear integration. The hosted alternatives run an agent SDK loop you can't see or take over, and bill through an API key. This runs the actual `claude` CLI with [Remote Control](https://code.claude.com/docs/en/remote-control) on, so the work happens on your machine, on your Claude subscription, in a session you can jump into mid-task.
What happens when you assign an issue
- Linear sends an agent session webhook to a small relay on your machine.
- The relay makes a git worktree for the issue, on branch
claude/. - It starts
claudeintmuxwith the issue as the prompt and Remote Control on. - The
claude.ai/codelink is posted back on the Linear issue. - Replying on the issue types into the running session.
- Claude commits, pushes and opens a PR, then comments the link.
Pick the model per issue with `[model=sonnet]` in the description, or a `haiku` / `sonnet` / `opus` label. Sessions run with `--dangerously-skip-permissions`, since nobody is at the terminal to answer a prompt — `DISALLOWED_TOOLS` is what still holds. Read [Worth knowing](#worth-knowing) before you point this at a repo that matters.
Setup
Paste this into Claude Code from the repo you want the agent to work in:
Set this up for me: https://github.com/julian-pineiro/linear-claude-agent/blob/main/SETUP.md
Read that file and follow it step by step. Ask me for anything only I can do.
It takes about ten minutes, most of it yours: creating an ngrok account and a Linear OAuth app. To do it by hand, follow [SETUP.md](SETUP.md) yourself.
Requirements
- macOS (Linux works except the launchd services — use systemd user units)
- Claude Code, logged in with
claude auth loginon a Pro or Max plan. Remote Control does not work with an API key or a `setu
Related Skills
Entry
Load a Linear issue into context and enter plan mode for a focused session
Linear Continue Debugging
Continue debugging Linear issue: $ARGUMENTS
Linear Skills
A lightweight Claude Code skill for fetching Linear issue details without the context overhead of the full Lin
Label
Add or confirm a label on a Linear issue (bug/chore/feature)
Deal Review
Run a structured deal strategy session — assess deal health, identify risks, build a win plan, and assign next
Tmux Master
Tmux master skill to allow claude code and codex to play the orchaster role to fork tmux session reversed for
Related Agents
Program Manager
Proactively use for comprehensive Linear project management, release planning, dependency analysis, and projec
Linear Task
Fetch context from a Linear issue before starting work. Reads issue details, comments, and related issues.
Integration Manager
Cross-platform synchronization specialist for GitHub, Linear, and other tools. MUST BE USED for issue tracking