Tests for the changed code nothing covers. CloudAEye writes unit tests for uncovered code in an open pull request, including new test files where needed, and posts them as suggestions for you to review and run.
Generate unit tests for uncovered code in an open pull request, including new test files where needed. Posts the tests as review suggestions for you to review and run.
acme/orders-api #412
On the pull request
tests/billing/test_refund.pyNew file
import pytest
from billing.refund import apply_refund
def test_refund_cannot_exceed_remaining_balance(partly_refunded_order):
with pytest.raises(ValueError):
apply_refund(partly_refunded_order, amount=80)
def test_negative_refund_is_rejected(paid_order):
with pytest.raises(ValueError):
apply_refund(paid_order, amount=-5)Review the tests, apply them, and run them before you merge.
In your coding agent
/cloudaeye:add-tests #4126 tests proposed
| File | Suggestions | New |
|---|---|---|
tests/billing/test_refund.py |
5 | yes |
tests/billing/test_invoice.py |
1 |
1 posted as a one-click suggestion, 5 as comments.
A new test file is outside the PR's diff, so it goes up as a comment instead of a one-click suggestion.
The untested branch is a reviewer's second comment. CloudAEye finds the changed code no test reaches and proposes the tests, edge cases included.
A changed function is uncovered when no test reaches it. That is the same definition AI Slop's missing-tests signal uses, so the two agree.
When the test file does not exist yet, CloudAEye proposes the whole file, not just a few lines.
Tests arrive as review suggestions. Nothing changes until you apply them, and your working tree is never edited.
Tests use the framework your project already runs. CloudAEye detects it from your existing tests, and you can set it per language or per repository.
Claude Code
/cloudaeye:add-tests #412
Codex
$cloudaeye:add-tests #412
@cloudaeye /add_tests