Four hats. Four gates.
4hat is a quality framework for shipping software. Every change passes four gates, and each gate has one reviewer wearing one hat, one question to answer and one clear exit.
Tap a hat to fail its gate and see what happens to the release.
What each hat checks
Build gate
The hard hat asks: is it built properly?
- Lint and type checks pass
- Unit tests pass
- A second person reviewed the code
Exit: green build and one approving review.
Test gate
The detective's cap asks: does it work the way people use it?
- Integration and end-to-end tests pass
- Regression suite is clean
- Test coverage did not drop
Exit: no open critical bugs.
Security gate
The black hat asks: how could this be abused?
- Dependencies audited for known flaws
- No secrets or keys in the code
- Static analysis and permissions reviewed
Exit: no high-severity findings.
Release gate
The captain's cap asks: is it safe to put in front of users?
- Canary rollout is healthy
- Performance stays within budget
- Alerts are live and rollback is tested
Exit: rollback works and the dashboard is watched.
How the gates work
In order, fail fast
Cheap checks run first. A broken build never reaches the security scan.
The same bar for everyone
Hotfixes and big launches pass the same four gates. Only the size of the checks changes.
Automate, then add judgment
Machines run what can be scripted. People wear the hat for what needs a decision.
One owner per gate
Every gate has a named person who can say stop, and nobody can skip it silently.