code-review

Reviewing pending git diff changes requires comparing the current branch to the base branch and analyzing the complete diff with surrounding context to evaluate correctness, security, and test coverage, yielding evidence-backed findings grouped by severity.

Severity triage in a pull request code review groups evidence-backed findings into categories like Blocker, Concern, Suggestion, and Praise, citing specific file and line locations to help reviewers focus on defects and contract-breaking changes instead of stylistic nitpicks.

Running type-checkers and linters during a pull request review involves executing cheap type-checkers, linters, and tests when feasible against the changed behavior. This verifies coverage and acts as a quality gate before outputting a structured triage verdict.

Yes, this code review approach works for evaluating API and contract impacts across a full diff by reading the complete changes, checking surrounding code to understand invariants and call sites, and verifying that contract-breaking changes are identified with exact file and line evidence.

The best way to avoid stylistic nitpicks in static analysis and code review is to focus on defects, risks, and contract-breaking changes by producing a structured triage verdict that cites exact file and line evidence without restating the diff.

Code review not working to catch real defects often happens when reviewers lose focus on the full diff context. Applying evidence-based findings by checking surrounding code, call sites, and running cheap tests ensures actual defects are identified and grouped by severity.