The shop you secure (defence in a real app)

An ordinary system redrawn with every door an attacker could use marked on it. Every lesson in this track has walked through one of those doors. This one builds the building and puts a lock on each door itself.

Before you start. The Security Mindset gave you the threat model: a written list of the ways a system could be made to misbehave, walked door by door with a six-letter checklist called STRIDE. Authentication through Application Hardening each taught one kind of lock, Whose page is this? (the browser security model) showed which rules the browser enforces for you, and Review the fix (secure code review) had you find missing locks in other people’s code. A shop built to be broken (OWASP Juice Shop practice) let you walk through a shop left open on purpose. This lesson is the other side: a shop you secure, with a test that proves each lock is there. You need Python 3, which most laptops already have.

The idea in 30 seconds

A threat model is a written list of the ways an app could be abused, made by walking STRIDE at every door where outside input comes in.
This shop’s six doors each get one defence: an ownership check, parameterised SQL, output escaping, slow salted password hashing, a per-session form token, and an allowlist for addresses the shop fetches.
A defence is usually one line, and deleting it breaks nothing you can see, which is why each one needs its own test.
A regression test is kept forever so a closed hole cannot quietly reopen, and it is named for what it guarantees.
You trust a test only after you have watched it fail with its defence removed.
A good threat model names what it does not cover; this shop has no audit log and no rate limits.

Read the story · Skip to the numbers