Before your weekly delivery sync
Know what moved, what is stuck and what to do about it, with a Slack post ready to paste.
How it works
- 01
Copy the prompt
Fill in the [bracketed] placeholders or leave them: the assistant asks for what is missing.
- 02
Paste it into your AI app
Claude, ChatGPT, Cursor or your own app, connected to DevStats MCP. Set up DevStats MCP
- 03
Get the deliverable
slack-ready-post-and-3-actions.md
The prompt
I'm getting ready for my squad's weekly delivery sync. I want to walk in knowing what moved, what is stuck, and what we should do about it this week.
## Context
- Squad: [squad] (call List Squads if you need the exact name)
- This week: [last 7 days], compared with the 7 days before it
- Audience: the squad itself. They know the work, so skip definitions and be specific.
- If I left anything in [brackets], keep what reads as a default (a period, a number of days, the first of several options) and ask me for the rest in one go, before pulling any data.
- Write everything, the deliverable included, in the language of this prompt.
## Pull this data from DevStats (this week and the previous week, same filters)
1. Throughput — PRs opened vs merged
2. PR Cycle Time — total and the coding / pickup / review / deploy breakdown
3. Issue Cycle Time
4. Daily WIP — open branches and open tasks
5. Deploy — deploy count
6. Current snapshot: Open PRs and Aging Issues, with the oldest items and their age
7. Sprint Progress — only if the squad runs sprints
## How to analyze
- Compare week over week, but treat one week as a signal, not a trend. Say "up this week", never "getting worse".
- Opened > merged two weeks in a row means work is piling up in review or in draft. Check Open PRs to see where.
- Pickup time up with review time flat means PRs are waiting for someone to start, not for a long review. That is reviewer availability, usually one or two people.
- WIP up with throughput flat or down means too much started and not enough finished. That is context switching.
- Aging issues in the same project usually share one cause: a dependency, a missing decision, an environment. Name the cluster, not the individual tickets.
- Ignore moves under ~15% unless they continue last week's pattern.
- If a tool returns no data, say so and move on. Do not fill the gap.
## Deliver
1. A Slack post I can paste in the squad channel, under 150 words: what moved, what is stuck, what we are doing about it. Plain language, real numbers, one line per point. Name PRs and issues, not people.
2. Up to three actions for the week. Each one: the action, the evidence behind it, and who picks it up. Example: "Reviewers: two PRs open 6+ days in the API repo. Pick up or close today." In a quiet week, one action or none is the honest answer; do not fill the list.
3. One question I should ask the team that the data cannot answer.
What it returned
Download the exampleStop guessing. Start asking.
Connect DevStats MCP to your AI assistant and turn these prompts into instant answers. Copy a prompt and go.