15 posts.
Claude Code built the caseworker admin console for my apidays workshop demo and every check it ran passed - type check, lint, build, direct FGA queries. The first time I actually clicked the button, we found five separate bugs, one in each layer of the stack.
Part three of building the apidays workshop demo with Claude Code. Three separate times in one session I went back to the original brief and found real drift - and each time I described it to myself as though it were a fresh insight, rather than noticing I'd said the same thing before.
Verification level answers who a persona provably is, established once and true forever. Session assurance answers something else entirely - how strongly the current login just proved it - and it has no business being stored as a fact at all.
Part two of the apidays workshop series moves from planning to building. Claude reversed a hosting decision that turned out to be solving a problem that never existed, then quietly let a fourth infrastructure provider into a build I'd explicitly ruled out - twice.
An adversarial review of a conference workshop's build plan found a genuine structural flaw in its Auth0 FGA model, not a polish problem. The check it ran on a citizen's own record was trivially always true, which meant the demo's actual premise - authorising an AI agent acting on someone's behalf - was never being tested at all.
An adversarial review across two different AI models caught a real OpenFGA semantics bug in the model behind this series, and also produced confident, specific-sounding nonsense that didn't survive being checked. Both outcomes are worth writing up.
The model from this series has humans and businesses as principals, but not agents, and no way to represent a delegation that's been requested but not yet approved. Both turned out to need real additions, not workarounds.
Every delegation in this series so far has been bounded by time. This post covers the other kind: bounded by a dollar figure, or a dollar figure and a category together, checked fresh against every attempted spend rather than a stored running total.
A company registry needs three things a subscription service doesn't: a public entry anyone can search, a governance role distinct from day-to-day admin, and proof that whoever holds that role is who they say they are.
It's easy to claim an authorisation model is generic. It's more convincing to run the exact same model, unmodified, across media, healthcare, financial services, telco, retail and legal case management, and show the tests passing every time.
The delegated-access series treated the authenticated login as the whole security principal. It isn't quite. The same real person routinely holds more than one login, and an authorisation model that can't tell them apart ends up merging blast radii it shouldn't.
The previous post covered the CIBA approval flow for delegated access. This one covers what happens when the five-minute CIBA window closes before the approver responds - and why the answer reveals an important architectural distinction between the delivery mechanism and the approval itself.
The previous post covered how a patient proactively grants a carer access to their record. This post covers the inverse: a user requests access they do not yet have, and the account holder approves it in real time using Auth0's Client-Initiated Backchannel Authentication.
The first post in this series covered why separating the authenticated actor from the account they are acting within matters. This post covers how to build it: the data model, the integration points between Auth0, Auth0 FGA and a standard application data store, and the key flows from delegation grant through to acting on someone else's behalf.
Identity systems are very good at answering one question: who are you? What they are less good at is the follow-up: and who are you acting for? In this post I will look at why separating the authenticated user from the account or person they are acting within is important, what it enables across a range of industries, and how Auth0 and Auth0 FGA address the two complementary sides of the problem.