Merge request nudger
By Toolspoke
Lists the GitLab merge requests waiting too long for review each weekday, flags failing pipelines and posts them to Slack by reviewer.
Needs a Toolspoke account. You choose which project it installs into.
- GitLab + Slack
- Required integrations
- Runs on a schedule
- Weekdays at 10am by default
- Post one message
- What its last step does
- Some actions can delete
- 1 of 34 actions in its tools
On this page
The short version
- What it is
- A ready-made Toolspoke agent that posts the GitLab merge requests waiting too long for review, grouped by reviewer.
- What it reads
- Open, non-draft merge requests in the projects you chose, with their reviewers, approvals and pipelines.
- What it finds
- Merge requests past your waiting time, those with no reviewer, and those with a failing pipeline.
- What you get
- One Slack message each weekday morning, oldest first in each reviewer's group.
- How often
- Every weekday at 10 AM once you switch it on.
- What it never does
- Comment on, approve or merge a merge request.
- What you need
- A Toolspoke account, a GitLab access token, and a Slack workspace.
What Merge request nudger will never do
Merge request nudger is told never to comment on a merge request, never to approve or merge one, and never to @mention someone you did not map. It nudges in Slack, and the review stays with your team.
Merge request nudger will never
- Comment on a merge request.
- Approve, merge, close or reassign a merge request, or change its reviewers or labels.
- Open, edit or close an issue.
- @mention someone in Slack who was not mapped at install.
- Edit or delete a Slack message, or post more than once a run.
Why it will not comment in GitLab
A note on a merge request notifies everyone on the review, and any quick action in it, such as /assign, runs. A daily bot note becomes noise in every reviewer's inbox. The nudge goes to Slack instead.
Why it will not change the review
Who reviews, what gets approved and when it merges are the team's decisions. The agent lists the queue and changes nothing in it.
Why it will not guess a Slack handle
An @mention pages a person, and a wrong guess pages a stranger. Unless you map GitLab usernames to Slack names at install, it shows names without a mention.
These are written into Merge request nudger's instructions. They guide the agent, but they are not a permission. What it is allowed to touch is set by the access you grant.
Merge request nudger reads the queue and posts it. Who reviews, what gets approved and what merges stays with your team.
What Merge request nudger does
Merge request nudger is an AI agent that reads the open merge requests in your GitLab projects each weekday, finds the ones that have waited too long for a review, and posts one Slack message grouped by reviewer. A failing pipeline is marked, so nobody reviews a change that does not build.
Review queues grow quietly. A merge request opened on Tuesday afternoon sits behind a meeting, then a release, and on Friday its author is still waiting.
GitLab shows each reviewer their own list, if they look. Nobody sees the whole queue, or the merge requests that have no reviewer at all.
Merge request nudger posts the queue where the team already talks. Each reviewer sees their own merge requests, oldest first. Unassigned ones go to the top, and failing pipelines go back to their authors.
In practice, every weekday you have:
- Every open merge request that has waited past your limit
- Those with no reviewer, first
- Each reviewer's queue, oldest first, with how long each has waited
- Merge requests with a failing pipeline, sent back to their authors
How it works
Merge request nudger works in four steps each weekday: it lists the open merge requests, keeps the ones waiting, checks their pipelines and posts one message. All of it comes from the agent's written instructions, which you can read and edit after you install it.
Step 1: List open merge requests
In each project you chose, drafts left out.
Step 2: Keep the ones waiting
Open longer than your limit, counting weekdays only.
Step 3: Check the pipeline
Mark a failing pipeline and name the job that broke.
Step 4: Post one message
Grouped by reviewer, oldest first.
It counts working days
By default a merge request is waiting after 24 hours, counting weekdays only. A merge request opened on Friday afternoon is not overdue on Monday morning.
It measures from when it was opened
GitLab does not record when a review was requested, so the wait runs from the merge request's created date. Drafts are left out, because their author has not asked for a review yet.
It sends failing builds back
If the pipeline failed, the merge request goes in a group for its author, with the first failed job's name and stage from list_pipeline_jobs. A reviewer's queue only holds what is ready to review.
It nudges in Slack only
It never comments on a merge request, because a GitLab note notifies everyone on the review and runs any quick action in it. The limits are listed under what it will never do.
What a review queue post looks like
A run of Merge request nudger posts one message to Slack. It leads with merge requests nobody has been asked to review, then each reviewer's queue, then failing pipelines.
Northwind
Channels
- general
- engineering
- eng-reviews
- releases
eng-reviews
7 merge requests waiting across 3 projects. Oldest: acme/api !482, 4 working days.
No reviewer (2)
- acme/web !1190 New billing settings page, by Leila Haddad. 3 days.
- acme/infra !77 Bump Postgres to 16, by Owen Price. 1 day.
By reviewer
- Mei Tanaka (2): acme/api !482 Rate limit per API key, by Sam Okoro, 4 days. acme/api !479 Retry failed webhooks, by Sam Okoro, 2 days.
- Pablo Ortiz (2): acme/web !1184 Fix date picker on Safari, by Leila Haddad, 2 days. acme/api !476 Paginate audit log, by Grace Liu, 1 day.
Failing pipeline, back to the author (1)
acme/web !1187 Upgrade to React 19, by Grace Liu. test:unit failed in stage test.
- No reviewer
Merge requests past the limit that nobody has been asked to review. They go first, because nobody else will notice them.
- By reviewer
Each reviewer's waiting merge requests, oldest first, with the author and how long each has waited in working days.
- Failing pipeline
Merge requests whose pipeline failed, with the first failed job. They are waiting on their author, not their reviewer.
When the queue is clear
If nothing has waited past the limit, it posts one line saying so. Merge requests already approved and waiting to be merged are listed in one line at the end.
If one project cannot be read, the post names it and carries on with the others. A merge request with no pipeline says no pipeline, rather than counting as passing.
Who uses Merge request nudger
Merge request nudger suits an engineering team that reviews in GitLab and talks in Slack, and where review time is the slow part of shipping. The examples use made-up names.
A tech lead keeping the queue moving
Mei Tanaka reads the no reviewer group first and assigns each one before stand-up.
A reviewer starting the day
Pablo Ortiz finds his name and starts with the oldest of his two.
An author waiting on a review
Grace Liu sees her upgrade in the failing group and fixes the unit test before anyone wastes a review on it.
An engineering manager watching cycle time
Owen Price watches the first line each morning. When the oldest wait passes a week, it is a staffing question.
What Merge request nudger connects to
Merge request nudger connects to two tools: it reads GitLab and posts one message to Slack. You connect each one yourself, and Toolspoke stores the connection encrypted, so the agent never sees the credential.
GitLab
RequiredBrowse projects, triage issues, review merge requests, and check pipelines.
Authentication: Personal access token
Slack
RequiredSearch conversations and send messages to your workspace.
Authentication: Bot token
| Tool | What Merge request nudger does | What it cannot do |
|---|---|---|
| GitLab | Lists and reads merge requests and the jobs of their pipelines (list_merge_requests, get_merge_request, list_pipeline_jobs). | Merge, approve or delete anything: the connector has no such action. With a read_api token it cannot comment, or open or edit an issue, either. |
| Slack | Posts one message to the channel you chose (send_message). | Delete messages: the connector has no delete action. It can edit its own messages and schedule messages, and the agent is told to do neither. |
That the agent cannot merge is a property of the connector, which has no merge action. That it cannot comment is a lock if you connect GitLab with a read_api token. That it posts one message in Slack is an instruction to the agent, not a lock. If you want a lock, a project can switch individual actions off. The access control guide explains how.
The agent reads the projects the token can reach. A project or group access token limits it to those projects only.
When Merge request nudger runs
Merge request nudger runs every weekday at 10 AM, but only after you switch its schedule on. It arrives switched off, so nothing runs and nothing is charged until you decide.
- Each run reads every open, non-draft merge request in the projects you chose.
- It keeps those past your waiting time, counting weekdays only.
- It posts one message to the Slack channel you chose.
- If nothing is waiting, it posts one line saying so.
Each run uses model credit from your workspace balance, so it makes one run each weekday. The Toolspoke tool calls themselves are free.
Your copy is an ordinary agent in your project. You can pause its schedule there, run it once on demand, change the time and edit its instructions. The agent's page lists recent runs and their cost.
How to set up Merge request nudger
To start nudging, add Merge request nudger to a Toolspoke project, connect GitLab and Slack, choose your projects and switch on the schedule. It takes a few minutes, and nothing runs until the last step.
Step 1: Create a Toolspoke account or sign in
Choose Use this agent. If you are new, you will be asked to sign up first, and you come straight back to this agent afterwards.
Step 2: Add it to a project
Pick the project it should live in. You get your own editable copy, and you can see the two tools it brings before anything connects.
Step 3: Connect GitLab with read_api
Create a personal, project or group access token with the read_api scope and paste it. For self-managed GitLab, add your instance URL.
Step 4: Choose projects and the waiting time
List the projects by full path, such as acme/api, and how long a merge request may wait. The default is 24 working hours.
Step 5: Connect Slack
Sign in to Slack and invite the bot to the channel the queue should post in. Optionally map GitLab usernames to Slack names.
Step 6: Switch on the schedule
The schedule arrives switched off. Run it once, check the queue against GitLab, then switch it on.
Make Merge request nudger fit your team
Merge request nudger's instructions are plain English and you can edit them after you install it. Installing copies the agent, so your edits stay with your copy. Open the agent in Toolspoke and tell the builder what to change, and it rewrites the instructions.
Paste any sentence below into the builder chat, or say the same thing in your own words.
Count a merge request as waiting after 8 working hours instead of 24.
Only include merge requests that target main.
Leave out merge requests labelled do-not-review.
Mention each reviewer by their Slack handle, using the mapping I gave at install.
Only names you mapped get a mention. Everyone else is still shown by name.
Also run at 3 PM on weekdays.
This changes when the agent wakes up, so the builder edits its schedule, not the instructions.
None of these edits needs more access or touches the Never rules. They change what Merge request nudger is told, not what the connections allow.
How it compares to other options
Merge request nudger beats the alternatives when reviews stall because nobody sees the whole queue, and loses to them when you want the review itself automated.
| Compared on | Merge request nudger | GitLab's to-do list | GitLab Slack notifications | Asking in stand-up |
|---|---|---|---|---|
| Shows the whole queue | Yes, every project | Your own only | As events happen | If people remember |
| Finds merge requests with no reviewer | Yes, first | No | No | Sometimes |
| Only nags after a wait | Yes, in working days | No | No, every event | Daily |
| Sends failing builds to the author | Yes | No | Pipeline alerts | No |
| Cost | A run a weekday | Included | Included | Meeting time |
GitLab's own Slack notifications are the better choice when you want to hear about each event as it happens. Merge request nudger is a once-a-day view of what has been waiting, which is a different job.
A reviewer's GitLab to-do list shows their own queue. The nudger shows the whole team's, including the merge requests nobody was asked to review.
Frequently asked questions
Can Merge request nudger merge or approve anything?
No. The GitLab connector has no merge or approve action. With a read_api token it cannot comment or edit an issue either.
How does it decide a merge request is waiting?
It is open, not a draft, not yet approved, and was opened longer ago than your limit, counting weekdays only. GitLab does not record when a review was requested, so the wait runs from the created date.
Will it @mention people?
Only people you mapped from GitLab usernames to Slack names at install. Everyone else is shown by name, without a mention.
Does it work with self-managed GitLab?
Yes. Add your instance URL when you connect. It works with GitLab.com, self-managed and GitLab Dedicated.
Does it work with GitHub?
Not as shipped, because it reads GitLab. You can ask the agent builder for a version on GitHub. It needs its own instructions.
Is Merge request nudger free?
It has no price of its own, and every Toolspoke plan includes agents. Tool calls cost nothing, but each run draws on your usage balance. Pricing has the details.
Keep the review queue moving
Connect GitLab and Slack, choose your projects and switch on the schedule. The first queue lands the next weekday morning.