Skip to content
← Tools / Skills
Skill · .md packagelkshrk/10 trustMetrics-checked
Linear AI logo

Linear AI

Linear as the source of truth for agentic feature delivery.

Visit official site →

About

A development workflow kit that makes Linear the backbone of AI-assisted delivery: it teaches agents to clean up an issue, refine it into an implementation-ready plan, track progress in the issue description, implement in a dedicated worktree, run bounded review rounds, and close the issue only after the code lands in main.

The workflow is deliberately rigid — in-use labels, worktree-only implementation, five-round review caps — which keeps agents honest but can feel heavy for small fixes that don't need the full ceremony.

What it can do

  • Linear issue refinement into implementation plans
  • Worktree-based implementation workflow
  • Bounded multi-round code review
  • Progress tracking inside Linear issues
  • Issue claiming protocol for multi-agent setups

Setup

  1. Install into Codex and Claude Code: npx skills add lkshrk/linear-ai --agent codex --agent claude-code
  2. Or install just the delivery workflow: npx skills add lkshrk/linear-ai --skill linear-deliver-feature --agent codex
  3. List available skills first if unsure: npx skills add lkshrk/linear-ai --list
  4. Set your Linear API key in the environment

Prerequisites

A Linear account with an API key; git worktrees supported by your setup

Use cases

codeautomationagents

Works with

Claude CodeCodex

Trust score — 6/10

/10 trustMetrics-checked

Scored 6/10 (metrics tier, metrics-v1). Evidence: Independent repo, README updated ~8 days before research; Install commands verified verbatim from the repo README; Rigid workflow by design: worktrees, in-use labels, five-round review caps; Single-maintainer project; longevity unproven. License: License not confirmed in sources read.

How the trust score is built

Full methodology →
Adoption · stars/downloads, log-scaled vs peers
6.0/10
Maintenance · recency, releases, issue load
7.0/10
Openness · license + docs
8.0/10
Manual review · sanity read + red-flag check
6.0/10
  • · Independent repo, README updated ~8 days before research
  • · Install commands verified verbatim from the repo README
  • · Rigid workflow by design: worktrees, in-use labels, five-round review caps
  • · Single-maintainer project; longevity unproven

Sources

Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.