After the pull request
It doesn't hand off. It stays on the code.
Opening a pull request is where most AI tools stop. CoderHelm keeps working on it until it is merged, and keeps watching the code after that.
- JiraTicket WEB-412 assigned to CoderHelm09:02
- coderhelmOpened draft #318 Send start date at checkout on
coderhelm/web-41209:14 - CICheck
unit-testsfailed: expected ISO date09:19 - coderhelmRead the logs, fixed the date format, pushed. All checks green09:26
- reviewer“Can we reuse the existing date helper?”11:40
- coderhelmSwitched to
toIsoDate(), replied with what changed11:47 - GitHubApproved and merged into main13:05
- coderhelmDeploy checks finished. No new failures since the merge13:31
Until the pull request is merged
Answers review comments
Request changes or ask a question on its pull request. It replies, pushes the fix and says what changed.
Fixes failing CI
When a check fails on its pull request, it reads the logs, fixes the cause and pushes again. It stops when the same failure repeats, and caps automatic pushes at eight. When only quarantined tests failed, it treats the run as passing instead.
Keeps the branch current
Resolves merge conflicts with the base branch before it works on feedback, so a pull request does not go stale while it waits.
Tells your failures from its own
When an end to end test fails for a reason the pull request did not cause, it runs it once more and says so, with the evidence.
After the merge
Checks the deploy
With auto merge on, it waits until the deploy checks finish and reports only failures that are new compared with before the merge. It never rolls back a shared branch.
Turns alerts into fixes
Production alarms routed to CoderHelm become analyzed pull requests in the right repository.
Keeps tests healthy
Flaky tests stop blocking builds, broken ones are never hidden, and fix pull requests come with proof.
Writes release notes
Tags releases after merges and writes notes your team can read, published to the GitHub release, Confluence or email.
FAQ
Does CoderHelm fix CI on pull requests it did not write?
It fixes failures on its own pull requests. On your team's pull requests the reviewer explains failures, and for end to end tests it says whether a failure is known flaky, a bad run or likely caused by the change.
Can it loop forever on a failing check?
No. It stops when the same failure happens twice and caps automatic pushes, then reports what it found.
Related
- Alerts
Route CloudWatch alarms and EventBridge events to CoderHelm through SNS. Matching alerts become analyzed pull requests in the right repository, and recoveries are paired with the alert that raised them.
- Flaky tests
CoderHelm reads the end to end results your CI produces, tells flaky tests from broken ones, keeps flaky tests from failing builds and opens fix pull requests that must pass twenty runs in a row.
- Code review
Add a label to any GitHub pull request and CoderHelm reviews it with the context of the whole repository, comments on exact lines with suggested fixes, rates risk and reviews again on every push.
Put CoderHelm on your next ticket.
CoderHelm is in open beta. Sign up and bring your own Anthropic API key.