Score every issue for AI-readiness and enrich specs automatically

Overview⁠

SpecBot analyzes every new issue in your Linear workspace and scores it 0–100 on how well-specified it is for AI coding agents to implement. It identifies missing acceptance criteria, repro steps, and technical context, then posts a structured comment with suggestions so your team can fix specs before they reach tools like Codex or Copilot.

How it works⁠

When a new issue is created, SpecBot receives a webhook and runs an AI analysis against the issue title, description, labels, and priority. It evaluates completeness across seven dimensions: clear title, description with context, acceptance criteria, steps to reproduce, scope boundaries, technical hints, and priority/labels. Each dimension is scored and combined into a single Codex-readiness rating from 0 to 100.

SpecBot then posts a comment on the issue with the score, a list of what's missing, and suggested content for each gap. For bugs, it drafts repro steps. For features, it generates acceptance criteria. For all issues, it flags when scope or technical context is absent. Teams can review the suggestions and incorporate what's useful.

The integration supports three modes: suggest-only (comments with recommendations), auto-apply (applies high-confidence suggestions directly), and stale detection (flags issues with no activity after a configurable threshold). All modes can be toggled from the SpecBot dashboard.

Configure⁠

Install SpecBot by clicking "Add to Linear" on the integration page or visiting the install link. You'll be redirected to Linear's OAuth flow to authorize SpecBot for your workspace. You must be a workspace admin to install. After authorization, SpecBot creates a webhook on your first team and begins monitoring new issues immediately. Configure settings at https://specbot.dasgroupllc.com/dashboard — toggle auto-commenting, auto-apply, and stale issue detection. No additional setup is required.