Automation or manual work

[sudoUgando]

Do you rely on automation (CI/CD, linters), or prefer manual review?

[EimanTahir027]

Do you rely on automation (CI/CD, linters), or prefer manual review?

Treat it as a false choice—you want both, but for different jobs.

What automation is best at

Automation (CI/CD, linters, tests) is your first line of defense:

Catches syntax errors, formatting, type issues (e.g., flake8, black, mypy)
Runs tests on every commit via pipelines like GitHub Actions or GitLab CI/CD
Prevents broken code from being merged
Enforces consistency across the codebase

In a solid setup, a PR shouldn’t even be reviewable if CI is failing.

What manual review is best at

Humans handle what automation can’t:

Code clarity and maintainability
Architectural decisions
Edge cases and unintended side effects
Naming, readability, and “does this make sense?”

A linter won’t tell you if your design is messy or your abstraction is wrong.

How good teams combine them

A common workflow:

Developer opens PR
CI runs automatically (tests, lint, type checks)
If CI passes → humans review logic/design
Only then merge