Sample deliverable

PR review and bug rescue report.

This sample shows the handoff format for a fixed-scope code review, failing-test triage, or one-flow rescue job. It is designed for small teams that need a concrete engineering decision, not a vague list of style opinions.

Before

The example project has a pull request that claims to fix a broken import flow. CI is red, the reviewer is not sure whether the patch is safe, and the demo owner needs a yes/no decision before merging.

Observed failure: `pytest tests/test_import_flow.py -q` fails on duplicate rows.
User impact: CSV imports can silently skip valid customer records.
Scope: Review the PR, isolate the defect, and propose a small fix.
Output: Findings, patch plan, verification log, and residual risk.

Review findings

High risk

Duplicate detection runs before normalization.

The PR compares raw rows before trimming account IDs and lower- casing email values. That means two records that should collapse into one customer can pass through as separate rows.

Medium risk

The test covers the happy path only.

The added test proves a clean import, but it does not cover whitespace, duplicate casing, empty optional fields, or malformed dates. The bug can return after a refactor without a failing signal.

Medium risk

The error message hides the bad row.

The current exception reports that import failed but does not include the row number or normalized key, making support triage slower than necessary.

Patch strategy

Normalize row keys before duplicate detection.
Add regression tests for whitespace, casing, and repeated customer IDs.
Keep the patch inside the import module and avoid unrelated cleanup.
Add a concise error message that identifies the row and normalized key.
Run targeted tests first, then the repo's standard lint/test gate.

Verification log sample

Reproduce: pytest tests/test_import_flow.py::test_duplicate_customer_rows -q -> failed before patch
Targeted: pytest tests/test_import_flow.py -q -> 9 passed
Lint: ruff check app/imports.py tests/test_import_flow.py -> passed
Format: ruff format --check app/imports.py tests/test_import_flow.py -> passed
Diff check: git diff --check -> clean
Not tested: live production database, private customer export, billing workflow

Client handoff

A paid triage ends with a review comment or short report. A rescue sprint ends with a patch, focused tests, and the same verification evidence. The goal is to make the next maintainer decision obvious: merge, request changes, or reject the risky direction.

Findings ordered by severity
Small patch plan tied to the failing behavior
Commands run and exact result
Known gaps that need client-side credentials or production data

Safe intake brief

Send a sanitized repo or PR link, the failing command, the expected behavior, and any deadline. Do not send credentials, customer data, payment details, private keys, or requests to bypass platform rules.

Need a review like this?

Open a quote request with the failing command or pull request link. I can quote a review-only triage or a small fix-and-proof rescue.

Request a quote AI app sample Back to services