Skip to content

PR Review Summariser

Example prompt: "Every morning, list all open pull requests on our GitHub repo, flag any that have been open more than 3 days, and post a review digest to #engineering on Slack."

How to automate PR review summaries with GloriaMundo

The Problem

Open pull requests that sit without review slow down the entire team. Code goes stale, merge conflicts accumulate, and the author moves on to other work, losing context on their own changes. Most teams rely on GitHub's notification emails to prompt reviews, but those get buried alongside every other notification. Without a clear daily picture of what is open and what is ageing, PRs can linger for days before anyone notices. Engineering leads end up manually checking the PR list each morning and chasing reviewers individually — a repetitive task that should not require human effort.

How GloriaMundo Solves It

We build a scheduled workflow that runs every morning. An integration step queries your GitHub repository for all open pull requests, pulling the title, author, creation date, assigned reviewers, number of comments, and CI status. An LLM step analyses the list and produces a structured digest: a summary of each PR (what it does, based on the title and description), how long it has been open, who is assigned to review it, and whether CI checks are passing. PRs that have been open longer than your threshold (say, 3 days) are flagged as stale with a note on how many days they have been waiting. An integration step posts the formatted digest to your engineering Slack channel so the team sees it first thing in the morning. Glass Box preview shows you the full digest before it is posted, so you can adjust the staleness threshold or the level of detail.

Example Workflow Steps

  1. Trigger (scheduled): Runs every weekday morning at 9:00 AM.
  2. Step 1 (integration): Query GitHub for all open pull requests — title, author, creation date, reviewers, comment count, and CI status.
  3. Step 2 (llm): Summarise each PR, calculate how long it has been open, flag any that exceed the staleness threshold, and format the results into a structured digest.
  4. Step 3 (integration): Post the formatted PR review digest to #engineering on Slack, with stale PRs highlighted at the top.

Integrations Used

  • GitHub — source of open pull request data including metadata, reviewers, and CI status
  • Slack — delivers the daily review digest to the engineering team

Who This Is For

Engineering leads, tech leads, and development teams who want to keep pull requests moving and ensure nothing sits unreviewed for too long. Particularly useful for teams of 5-20 engineers where multiple PRs are open at any given time and review bottlenecks are a recurring problem.

Time & Cost Saved

Manually checking the PR list, assessing which ones need attention, and chasing reviewers takes 15-30 minutes per day for an engineering lead. Over a month, that is 5-10 hours of administrative work. This workflow delivers a ready-made digest every morning, ensuring stale PRs are surfaced before they become a problem. The indirect savings are larger: faster reviews mean shorter cycle times, fewer merge conflicts, and less context-switching for PR authors. The workflow uses integration and LLM steps, costing a few credits per daily run.