_Follow along with this video:_ --- ::image{src='/security-section-3/12-protocol-tests/protocol-tests1.png' style='width: 100%; height: auto;'} As security researchers our job is to ultimately do what's necessary to make a protocol more secure. While we've thoroughly examined everything within scope of `PasswordStore` there can be some value in expanding our recon. Test suites should be an expectation of any protocol serious about security, assuring adequate test coverage will be valuable in a `private audit`. ## Testing and Coverage Anyone at this stage of the course should be familiar with how to check the `test coverage` of a repo. ```bash forge build forge test ``` The above will run all current tests, to check `coverage` we'll use: ```bash forge coverage ``` ::image{src='/security-section-3/12-protocol-tests/protocol-tests2.png' style='width: 100%; height: auto;'} Wow! Our coverage looks great...right? It's important to note that coverage may be a vanity metric and not truly representative of what's being tested for. If we look closely at the tests included, we can see the a major vulnerability we found (`Access Control`) wasn't tested for at all. ```js function test_owner_can_set_password() public { vm.startPrank(owner); string memory expectedPassword = "myNewPassword"; passwordStore.setPassword(expectedPassword); string memory actualPassword = passwordStore.getPassword(); assertEq(actualPassword, expectedPassword); } function test_non_owner_reading_password_reverts() public { vm.startPrank(address(1)); vm.expectRevert(PasswordStore.PasswordStore__NotOwner.selector); passwordStore.getPassword(); } ``` In addition to the above, tests aren't going to catch problems with documentation, or erroneous business logic. It's important not to assume things are fine because our framework tells us so. ### Wrap Up We're really progressing through this process well and we're ready to write a report for each of our findings. We'll cover this in our next lesson!