Skip to content

Safe dependency update pull requests

Finds outdated packages on a schedule, applies only small updates that pass your tests, and opens a pull request. Major updates are listed for a person.

EngineeringWorks with Claude Code · Cursor+1Community example
Build your promptSee requirements

What it does

  • Reads your package files and checks for newer versions
  • Sorts each update by risk: patch, minor or major
  • Applies patch and minor updates, then runs your tests
  • Rolls everything back if a test fails
  • Opens a pull request with a summary of what changed

Your approval boundaries

Creates drafts

These actions always wait for your yes:

  • Applying any major update
  • Merging the pull request
  • Touching packages on your exclude list

You decide what happens next.

More about this workflow

This setup is for developers and small engineering teams who want fewer chores. It takes a repeat chore off your plate and hands back something you can check quickly.

First, it reads your package files and checks for newer versions. Next, it sorts each update by risk: patch, minor or major. Then it applies patch and minor updates, then runs your tests. After that, it rolls everything back if a test fails. Finally, it opens a pull request with a summary of what changed. It works with your git host, CI and package registry, each on its own login with limited access.

You get the prep work done and waiting for you. The agent stops and asks for your yes before applying any major update, merging the pull request and touching packages on your exclude list, and says plainly if a step cannot be undone.

“They are never applied unless you ask.”
AlanRasch AlanRasch on GitHub · Oct 9, 2026
Before you startGit host · CI accessScoped permissionsClaude Code or Cursor
Explore required tools and model recommendations

Tools and access

  • GitHub

    Handles the git host and ci part of this job. Grant a fine-grained token, read code and write comments.

    API required: Yes. Needs a GitHub App or a fine-grained personal access token scoped to the repos it needs.

    Sign up: Create a GitHub account

  • Package registry

    Handles the package registry part of this job. Grant public read access.

    API required: No. Public registries can be read without an account.

    No sign-up needed

Official links only. Not sponsored, no affiliate links.

Model recommendations

Best pick

Claude Opus 5.5

Its maker points to it for long running coding work.

$4 per million input tokens, $20 per million output tokens.

Anthropic docs as of October 11, 2026

Cheaper backup

Claude Sonnet 5.5

Half the price per token and still strong for reviews and docs.

$2 per million input tokens, $10 per million output tokens.

Anthropic docs as of October 11, 2026

Our own picks. Not sponsored, no affiliate links.

Build your prompt

Add your details, then copy the prompt into your tool.

1Your details

Fill these in and the prompt updates as you type.

weekly, nightly

2Your prompt

4 of 4 still empty
Fill these in before you run it: [YOUR REPO] [YOUR TEST COMMAND] [PACKAGES TO NEVER TOUCH] [HOW OFTEN IT RUNS: weekly, nightly] You are the Safe dependency update pull requests agent. Finds outdated packages on a schedule, applies only small updates that pass your tests, and opens a pull request. Major updates are listed for a person. Use only the details filled in above. If any [BRACKETED] item is still blank or unclear, stop and ask for it first. Do not guess. What you do on your own: - Reads your package files and checks for newer versions - Sorts each update by risk: patch, minor or major - Applies patch and minor updates, then runs your tests - Rolls everything back if a test fails - Opens a pull request with a summary of what changed What you must never do without a yes from a person: - Applying any major update - Merging the pull request - Touching packages on your exclude list Rules: - Before any step on the list above, stop and ask. Say in plain words what you want to do, who it affects, and whether it can be undone. - If something cannot be undone, say so first, in one short line. - Write drafts, not final actions. A person sends, pays, posts, or deletes. - Keep a short log of every action you took and every source you read. - If you are unsure, ask instead of guessing.
View run instructionsPaste into: Claude Code, Cursor, OpenAI Agents SDK

Your answers stay in this browser only.

Run your workflow

Connect your tools with only the access this setup needs, then follow these steps.

Steps

Do these in order. Test on old data before it touches live work.

  1. 1Connect GitHub with its own login, not yours. Grant a fine-grained token, read code and write comments.
  2. 2Set up Package registry. Keep it to public read access.
  3. 3Paste the prompt where your harness takes instructions. See the note under the Copy button.
  4. 4Do a dry run on old or test data first. Check every draft and every stop point by hand before you let it near live work.
  5. 5Read its log after the first real run. Only widen access once a week of runs looks clean.

Safeguards and trust score

The score describes this example setup’s access, approval boundaries, ability to undo changes, and scope.

Access24 / 30

Writes drafts and notes inside your tools, nothing leaves on its own.

Human approval30 / 30

Waits for a yes on: applying any major update, merging the pull request, touching packages on your exclude list.

Undo and logs14 / 20

Proposes changes to live records, so the setup needs a change log and a way back.

Scope16 / 20

Runs on its own login across 3 tools. Fewer connections means a smaller blast radius.

How to raise it

  • Undo and logs: Stop it from editing live records. Have it write the change as a proposal you apply, and keep a log of every change with a way to roll it back +3
  • Scope: Cut it down from 3 connected tools to the 2 it needs most, each with its own limited login instead of your full account +2

Do all of these to reach 89, up from 84.

Want help getting there? Get an AI Trust Audit from Pixeldesigns

Reviewed by Trustcase (example setup). How scoring works

Embed badge

Show this trust score on your site or README. It links back here.

Trustcase trust score badge
Markdown
[![Trustcase trust score for Safe dependency update pull requests](https://trustcase.dev/badge/safe-dependency-update-prs.svg)](https://trustcase.dev/use-cases/safe-dependency-update-prs)
HTML
<a href="https://trustcase.dev/use-cases/safe-dependency-update-prs"><img src="https://trustcase.dev/badge/safe-dependency-update-prs.svg" alt="Trustcase trust score for Safe dependency update pull requests" height="20"></a>

More engineering workflows

  • Overnight safe task picker

    Each night it looks over the day's unfinished work and picks 3 to 5 coding tasks that are safe to run overnight. It only writes a report with a ready prompt for each one.

  • Upstream changelog impact triage

    Reads each new release of a tool your team depends on and lists which of your internal rules, skills and docs now say something wrong. It opens one issue with a follow up task per file.

  • Coding agent session journal

    Writes one note a day on what your coding agent sessions did, decided and learned, then rolls the week into a review. It reads transcripts already on disk and only runs when you ask.

  • Morning ticket dashboard with tick to approve queue

    Every morning an agent refreshes a one screen dashboard in your notes from the issue tracker, your workstream notes and inbox. It reads only, proposes board changes as checkboxes and drafts first pass PR reviews that are never posted.