eBPF Continuous Profiling released

AutoMerge · Beta

Approve the safe pull requests. Keep people on the rest.

AutoMerge scores each pull request and approves the safe ones. You start it in Dry run, where it only posts its decision as a comment until you switch it on.

Start for free or Book a demo

Routed to a human
Dash0Botcommented 2 days ago • edited

Darkplane AutoMerge

Not approved38.9 / 100
threshold 70
Correctness×0.4055
Test coverage×0.2522
Conventions×0.2024
Risk×0.1544

2 criteria scored below their required floors, which reject the run whatever the total.


Evidence

Correctness scored 55, below its required floor of 60.

The retry loop resets its backoff on every attempt, so a failing endpoint is retried without any delay.

Test coverage scored 22, below its required floor of 50.

The new retry path has no tests. The existing suite only exercises the success case.

Conventions scored 24.

Errors are logged and dropped instead of wrapped and returned, which the rest of the package avoids.

Risk scored 44.

The change sits on the path of every export, but stays behind the existing feature flag.

Approved
Dash0Botcommented 2 days ago • edited

Darkplane AutoMerge

Approved82.65 / 100
threshold 70
Correctness×0.4080
Test coverage×0.2587
Conventions×0.2092
Risk×0.1570

Evidence

Correctness scored 80.

The rename is applied consistently, including the one call site that builds the key by hand.

Test coverage scored 87.

Unit tests cover the new label and the migration of the old one.

Conventions scored 92.

Naming and error handling follow the surrounding package.

Risk scored 70.

It touches the query builder, but only the label text changes, not the query it produces.

Your agents write faster than anyone can review.

Coding agents made a pull request cheap to write. Reviewing one did not get cheaper, so the queue grows, including the routine changes nobody needed to read line by line. AutoMerge takes that routine share off your reviewers and routes everything else to a person.

01

Pins the commit

Every evaluation runs on one fixed commit. A new push means a new evaluation, and approvals on older commits are dismissed.

02

Holds what you protect

A pull request that touches a protected path goes straight to a person, without being scored.

03

Scores the change

One AI evaluator per criterion reads the change and scores it from 0 to 100, with a rationale you can read.

04

Approves or holds

It approves only when the combined score clears your threshold and no criterion falls below its floor. Everything else waits for a reviewer.

Every merge is a risk decision.

A good reviewer weighs the change, the person behind it, and how much risk the system can take right now. AutoMerge starts from these three and scores them the way you decide, so the same change can pass on a calm day and wait for a person during an incident.

The change

How much could go wrong?

Blast radius, evidence quality and change footprint score how far a failure could reach, whether there is proof the change works, and how routine it is.

The author

Do we trust who wrote it?

AutoMerge reads the author's track record: merged and later-reverted pull requests, and how much of the touched code they wrote.

The system

Can we afford it right now?

AutoMerge reads your system's health from Dash0. When services are healthy, a change scores higher. During an incident, the same change scores lower and waits for a person.

Your rules

Make it your own

Change any criterion's weight and prompt, add criteria for whatever your team cares about, or set a floor that rejects a pull request on its own, whatever the combined score.

Same change, different day

System health
Author
Pull request #482Approved
74/ 100
threshold 70
Change
74
Author
±0
System
±0

Because the system is healthy and the author has some history here, the change scores 74, clearing the threshold of 70.

Illustrative values.

Autonomy you grant one step at a time.

You start each repository in Dry run, where AutoMerge comments what it would do and changes nothing. Move a repository up once its decisions match your reviewers' decisions, or give more autonomy only to the paths that have earned it.

  1. 00Off

    Nothing runs

    Takes no action on this repository.

    • Score & comment
    • Approve
    • Merge
  2. 01Dry runStart here

    Watch it decide

    Comments its decision without approving or merging.

    • Score & comment
    • Approve
    • Merge
  3. 02AutoApprove

    Approve, a person merges

    Approves qualifying pull requests. Never merges.

    • Score & comment
    • Approve
    • Merge
  4. 03AutoMerge

    Approve and merge

    Approves and merges qualifying pull requests.

    • Score & comment
    • Approve
    • Merge

Check its decisions before you trust them.

Every decision, score and rationale lands in Dash0 next to what your reviewers actually did, so you know when a repository is ready for more autonomy.

Agreement

Graded against your reviewers

Each decision is graded as a correct approval, an incorrect approval, a correct hold or an unnecessary hold, so you can see where the bot and your team disagree.

Testing

Replay before you change

Test a new threshold, prompt or criterion against a fixed set of past pull requests and compare it with what is running. Tests never change what is running.

Feedback

Was this decision right?

Reviewers mark a decision right or wrong, with a reason, from the pull request in Dash0 or from the bot's comment on GitHub.

Built to fail closed.

You can only rely on an approval agent if you can predict when it will hold back. If an evaluator fails, a path is protected or a score falls short, AutoMerge leaves the pull request to a person.

Path rules

Protect what matters most

Mark migrations, infrastructure or billing as Protected and they always go to a person. A pull request runs at the least autonomous level of any file it touches, and renamed files are traced to where they came from.

Fails closed

When in doubt, it holds

If an evaluator fails, nothing is approved. In AutoApprove and AutoMerge, a commit whose CI is failing is never approved.

Untrusted input

A pull request can't rewrite its own rules

Rules live in Dash0 and steering files are read from the base branch. Titles and descriptions are treated as signals, not proof.

People decide

Advisory, never in the way

Its GitHub check is advisory and blocks nothing unless you make it required. It never re-approves a pull request a reviewer dismissed, and only admins can change a repository's configuration.

Reads what a reviewer would read.

Each criterion chooses its own read-only tools. Evaluators read tracked files at the evaluated commit and never run your code.

Codebase

The code around the change

Read files, grep, list files, git log, blame and the diff against the merge base.

Pull request

The change and its signals

Metadata, changed files, CI check runs, reviews and comments, and commit signatures.

Track record

Who wrote it, and how it went

The author's familiarity with the files, their delivery record, AI review agent activity, and the repository's evaluation history.

Coding sessions

How the change was built

Which harness, tools and skills produced the branch, read from AI SDLC Insights.

Start in Dry run. Switch it on when it has earned it.

Connect GitHub, pick your repositories, and see what AutoMerge would have approved before it approves anything.

Book a demo