The week in engineering, written for leadership.
Weekly eng update is an AI agent that writes your weekly engineering status report from GitHub. It counts the week's merged pull requests, releases and closed issues, groups the work into a few themes, lists what's in progress and the risks, saves the report as a new Notion page and posts a short summary with the link.
- 16:00Schedule fired
- 16:0064 PRs merged, 12 by bots
- 16:019 releases, by component
- 16:02Saved report in Notion
- 16:03Posted summary · 2 risks
Why do engineering status reports take so long?
Because the facts are spread across a hundred pull requests, a dozen releases and the issue tracker, and someone has to read them all on a Friday. The counts get done by hand, the doc and the message disagree, and the risks that matter most depend on who happened to write it.
Without it
- Friday afternoon spent scrolling a hundred pull requests.
- Numbers that differ between the doc and the message.
- Releases counted by hand, per component.
- Risks that only one person knows about.
With Weekly eng update
- The week's work grouped into a few themes.
- Merged, bot and release counts that match everywhere.
- Open work and up to five risks, with evidence.
- One Notion page and one message, every Friday.
What goes in an engineering status report?
Four answers for leadership every week: what shipped, what was released, what's in progress and what could go wrong.
What shipped
Merged pull requests grouped into four to six themes, each with a line of impact and linked examples.
Releases
Every release published this week, counted by component, with the first and last version.
In progress
Open pull requests and active issues, grouped, so leadership sees what's next.
Risks
Up to five open problems read in detail, with evidence, and inferences marked as inferences.
How does an AI engineering status report work?
Every run follows the same five steps. GitHub is only read; the report is written once, as a new Notion page, then summarised in Slack.
- 1
Counts what merged
Monday 00:00 in your time zone to now. Bot pull requests are counted apart.
GitHub - 2
Lists the releases
Releases published this week, by component, with the first and last version.
GitHub - 3
Reads open and closed work
Closed issues, open pull requests and up to five risks read in detail.
GitHub - 4
Saves the report
One new page under the page you chose, with the numbers counted once.
Notion - 5
Posts the summary
Under 120 words, the same numbers, and the page link.
Slack
- Fridays at 16:00, before the weekend
- @weekly-eng-update write this week's update
- Run when another tool calls its webhook, such as before a board meeting
From one sentence to Friday's report.
Describe the job, let Arclyx draft the agent, test it once on your real repository, create it, then read what lands in #eng-leads. The full report waits in Notion.
Paste the sentence. Arclyx picks the tools it needs and asks what it can't work out.
What should it do?
Every Friday at 4, write the weekly engineering update for leadership from this week's merged PRs, releases and closed issues in our GitHub repository: what shipped, what's in progress, and risks. Save it as a new Notion page under Engineering updates and post a short summary with the link to #eng-leads.
- Tools it can use
- GitHub
- Notion
- Slack
- Which Notion page should the reports go under?
- Engineering updates, in our leadership space
It writes the instructions, and you sign in to each tool once. Nothing runs until you say so.
- Reading your description
- Writing the instructions
- Picking the tools
Weekly eng update
- GitHub
- Notion
- Slack
Fridays, 16:00
Try it once as a dry run. It reads your real tools, and what it would post is shown to you, not posted.
- Counts what mergedGitHub · Search pull requests ×4
- Lists the releasesGitHub · List releases
- Reads open and closed workGitHub · List issues ×3
- Saves the reportNotion · Create page
- Posts the summarySlack · Post
Once it's created, the output lands where you said, with a source for every item.
Failed webhook deliveries now retry, and slow report queries moved off the main database.
Examples: #512, #527, #530.
↗ example-a.com/pull/512Nightly backups skip the largest workspace, and nobody is assigned to it.
Read in detail, with the evidence on the page.
↗ example-a.com/issues/433API 5 releases, web app 3, mobile app 1, each with its first and last version.
Only components with a release this week are named.
↗ example-a.com/releasesSearch indexing in two slices, the new billing page and a pending release.
Grouped by theme, with links.
↗ example-a.com/pulls
One report, four readers.
The same Friday page answers a different question for each person who opens it.
Engineering leaders
A report for leadership they only have to read and forward.
Executives
What shipped and what's at risk, without a meeting.
Product managers
Which themes moved this week, with links to the work.
Engineering managers
The week's numbers, counted the same way every Friday.
How do you write an engineering status report with AI?
Give an AI agent read access to GitHub and one page in Notion. Each week it counts the merged pull requests, releases and closed issues, groups the work into themes, reads the risks and writes the report once, so the page and the message carry the same numbers. In Arclyx that's five steps, starting with one sentence.
“Every Friday at 4, write the weekly engineering update for leadership from this week's merged PRs, releases and closed issues in our GitHub repository: what shipped, what's in progress, and risks. Save it as a new Notion page under Engineering updates and post a short summary with the link to #eng-leads.”
- 1
Describe the job
One sentence: which repository, which Notion page it goes under, and where the summary goes.
- 2
Connect GitHub and Notion
The agent gets GitHub tools that search and read, and a Notion tool that only creates new pages.
- 3
Name the parent page
Every report lands under that page, and the agent checks its title before writing.
- 4
Watch a dry run
Optional. The page is held and the summary shows a placeholder link, so you can read both first.
- 5
Schedule it
Fridays at 16:00, or @mention it before a leadership meeting.
Or start from one of these sentences
“On the 1st of each month, write last month's engineering report from our GitHub repository and save it under Monthly updates in Notion.”
“Every other Friday, write a one-page engineering summary for the board: what shipped, what slipped and the top three risks, saved in Notion.”
“Every Friday at 4, list this week's releases by component with their first and last version, and post them to #releases.”
“Every Monday at 9, read the open urgent issues in our repository and send me the top five risks with evidence.”
AI status report vs GitHub Pulse and a doc.
Writing it by hand and GitHub's Pulse page each cover part of the job. Here is what each one does, and doesn't, do.
| Written by hand | GitHub Pulse | Weekly eng update | |
|---|---|---|---|
| Merged pull requests | Scrolled and tallied | Listed for the period you pick | Counted and grouped into themes |
| Releases by component | Counted by hand | No | Yes, with first and last version |
| Risks with evidence | If someone remembers | No | Up to five, read in detail |
| Written for leadership | Yes, eventually | No: activity lists | Yes, in plain words |
| Where it lands | A doc and a message | A page in GitHub | A Notion page and one Slack message |
Source: GitHub docs: viewing a summary of repository activity (Pulse), checked 24 October 2026.
Is it safe to give an AI agent your GitHub and Notion?
It reads your repository, creates one page under the Notion page you choose and writes one message. Every run is recorded and has a spend ceiling by default.
Every run is recorded
Every step, in order
Every run records the tools it called, with a short preview of each result. Only the person who started a run can open its content.
Numbers counted once
Its instructions say to finish counting before writing, then reuse the same numbers in the page, the message and its reply.
What each run cost
Tokens and cost are recorded per run and per agent, next to your plan's limits.
A spend ceiling, on by default
Each run has a dollar ceiling unless you remove it. A run stops before it goes over it, and tells you.
It can
- Search and read pull requests, releases and issues
- Create new Notion pages, under the page you name
- Post to one Slack channel you name, message you, or email you
It can't
- Comment on, merge or close anything in GitHub
- Edit, move or delete a Notion page
- Email anyone but you
Engineering status reports, answered.
What goes in an engineering status report?
What shipped, what's in progress and what could go wrong, with the numbers behind it. Weekly eng update writes a two-to-four-line summary, the shipped work in four to six themes, releases by component, closed issues, open work and up to five risks, then the dates and caveats.
Where does the report go?
A new Notion page under the page you name, titled with the week's dates. It fetches that page first and stops if the title doesn't match. It never edits, moves or deletes an existing page. The Slack summary carries the link.
Does it list every pull request?
No. A busy week can have a hundred or more, so it groups them into four to six themes with a few linked examples each, and states the total merged and the exact count of bot pull requests.
How does it pick the risks?
From open urgent and high-priority issues and pull requests that look stuck. It reads up to five in detail, marks the rest as titles only, and labels anything it inferred, such as a pattern across several closed issues, as an inference.
How exact are the numbers?
Merged pull requests and releases are counted from GitHub's own lists for the week. Closed issues are an upper bound, because GitHub filters them by last update, not close date, and the report says so.
Does it change anything in GitHub?
No. It's given only GitHub tools that search, list and read. It can't comment on, merge, label or close anything.
What if Notion is unavailable?
It retries once. If the page still can't be created it doesn't post a link it doesn't have: it says the page failed and puts the update in the run's reply instead.
Do I need to write code?
No. Describe the update in one sentence. Arclyx drafts the agent, you sign in to GitHub, Notion and Slack once, and you can watch a dry run before you create it.
Agents that work well with it.
Build your weekly eng update.
One sentence in, a working agent out. Join the beta and we'll set you up.