Skip to content

agentik-os/linear-feedback-setup

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 

Repository files navigation

Linear Feedback Setup for Claude Code

One command to add a complete Linear feedback system to any Next.js project.

Install

curl -fsSL https://raw.githubusercontent.com/agentik-os/linear-feedback-setup/main/install.sh | bash

Usage

Open Claude Code in any Next.js project and run:

/linear-setup

That's it. Claude handles everything:

  1. Asks for your Linear API key
  2. Connects to Linear, finds/creates team + project + labels
  3. Installs dependencies (@linear/sdk, html2canvas-pro, @anthropic-ai/sdk, sonner)
  4. Creates the feedback widget component
  5. Creates API routes (authenticated + public)
  6. Configures MCP + Vercel env vars

What You Get

Feedback Widget (Dashboard)

  • Element targeting — click Feedback button, crosshair appears, click any element
  • Dual screenshots — global page (JPEG 0.6, max 1MB) + targeted element (JPEG 0.7, max 500KB)
  • Console error capture — last 10 console.error calls attached automatically
  • AI description improvement — "Improve with AI" button rewrites feedback for clarity
  • Auto-improve on submit — if user didn't manually improve, auto-runs before submit
  • Collapsible sections — screenshots and element info tucked away, expandable
  • Fullscreen preview — click any screenshot to view full-size
  • User Agent + URL — sent with every submission

Public Feedback Button (Marketing Pages)

  • Floating pill — bottom-left corner, same targeting flow
  • Honeypot anti-spam — hidden input field, bots get fake success
  • IP tracking — logged for abuse prevention
  • No auth required — works for anonymous visitors

Linear Integration

  • Issues created with proper labels (Bug, Feature, Improvement)
  • Source labels (Source: User Feedback vs Source: Public Feedback)
  • Screenshots uploaded via linear.fileUpload() as separate comments
  • Priority mapping (bugs = high, features = normal, questions = low)

Workflow

User submits feedback
    → Issue created in Backlog
    → AI agent works on it → moves to In Progress
    → Fix complete → moves to Review
    → Human validates → marks Done

Sub-commands

Command What it does
/linear-setup Full setup (all phases)
/linear-setup widget Only the feedback widget component
/linear-setup api Only the API routes
/linear-setup public Only the public feedback button + route
/linear-setup mcp Only MCP configuration

Requirements

  • Claude Code (CLI)
  • Next.js project with App Router
  • shadcn/ui (Dialog, Button, Textarea, Badge components)
  • Linear account with API key

Auth Support

The setup auto-detects your auth provider:

  • Clerk (@clerk/nextjs)
  • Better Auth (better-auth)
  • Auth.js (next-auth)
  • None (public-only mode)

Stack Compatibility

Stack Supported
Next.js (App Router) Yes
Clerk Yes
Better Auth Yes
Convex Yes
Supabase Yes
Prisma Yes
Vercel Yes (auto-configures env vars)

Update

Re-run the install script:

curl -fsSL https://raw.githubusercontent.com/agentik-os/linear-feedback-setup/main/install.sh | bash

License

MIT


Built by Agentik OS

About

One command to add a complete Linear feedback system to any Next.js project. Element targeting, dual screenshots, AI improvement, console errors — all via Claude Code.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors