Files
Ekklesia/core/tests/authorization-device.test.ts
ekklesiaandClaude Opus 5 80189c8ae2
CI / verify (pull_request) Successful in 40s
CI / infrastructure-smoke (pull_request) Successful in 21s
CI / migration-smoke (pull_request) Successful in 1m22s
feat: give display devices their own paired identity
ADR 0033 gave Ekklesia a display host and ADR 0034 gave it a login page.
Together, the only way to light a television in a foyer was to sign a
human being into it and leave that session running on a screen anyone can
walk up to: a person's credentials on a wall, a person's session as the
thing signage depends on, and the display going dark when they are
revoked or their session simply expires.

A screen is not a person. Ekklesia now recognizes three principal classes
— anonymous, authenticated human, and managed device — and a device is a
first-class principal with its own credential and its own explicit
permissions. It has no email, no password, and no row in the identity
tables, and it satisfies the `authenticated` baseline never, so every
route that settles for "somebody is signed in" stays closed to it.

Device permissions live in their own table because
ekklesia_core_grant.user_id is a non-null foreign key into the user table
and making it polymorphic would rewrite the platform's central
authorization record to serve one narrow case. That is a schema decision,
not a second decision point: the evaluator gained one branch, so a device
is allowed or denied by the same function, with the same deny-by-default,
as anyone else — with no owner path it could reach and no baseline that
stands in for a grant.

The six-digit code is not the credential. It is printed across a screen in
a public hallway, so it is treated as public: it identifies a request to
an administrator and is worth nothing alone. Collecting the credential
also requires a 256-bit claim secret held only in the requesting browser's
HttpOnly cookie, and approving requires the restricted core.devices.manage
permission. Without that split, reading a code off a wall would hand over
the screen's identity.

The permission a pairing asks for is read from the module's validated
display declaration, never from the request, so a browser cannot ask to be
granted core.devices.manage or another module's permission. Approval
grants exactly that one permission.

Credentials are stored only as SHA-256; the raw value exists in the
Set-Cookie that issues it and the device's own cookie jar, and is never
logged, returned in JSON, or placed in a URL. The cookie is HttpOnly, so
the module display code running in that same page cannot read it.
Revocation is a server-side column checked on every principal resolution,
and re-pairing mints a new identity so an old credential can never come
back. State transitions are conditional updates, so one code cannot be
approved twice or claimed into two credentials under a race.

/api/v1/principal still reports a device as anonymous: it is what the
Dashboard uses to decide whether a person is signed in. Displays read
their viewer from the descriptor instead, which already performed the
authorization — so displays are descriptor-first while the Dashboard stays
principal-first, and neither infers an identity by elimination.

Proven end to end against real PostgreSQL and real HTTP: an unpaired
screen shows a code, an owner approves it, the device opens the display
descriptor and a realtime stream, and revocation closes both while the
device still holds its cookie. An already-open SSE stream is bounded by
realtime's existing five-minute reconnect rather than closed instantly,
and that bound is documented rather than called immediate.

Co-Authored-By: Claude Opus 5 <[email protected]>
2026-08-23 23:41:03 -05:00

190 lines
6.8 KiB
TypeScript

import {
ANONYMOUS_PRINCIPAL,
createAuthenticatedPrincipal,
createDevicePrincipal,
parsePermissionId,
} from "@ekklesia/contracts";
import { describe, expect, it, vi } from "vitest";
import { createAuthorizationEvaluator } from "../src/authorization/evaluator.js";
/**
* Device authority, decided by the **same** evaluator that decides everything else.
*
* The point of these tests is mostly negative: a device must not acquire authority from being a
* principal, from being paired, from an `authenticated` baseline, or from platform ownership. The
* only thing that ever makes a device allowed is an explicit permission granted to that exact
* device (ADR 0036).
*/
const displayPermission = parsePermissionId("announcements.slideshow.view");
const otherPermission = parsePermissionId("calendar.display.view");
const deviceId = "3f1c8a4e-1d2b-4c3a-9e5f-6a7b8c9d0e1f";
function evaluator(options: { readonly devicePermissions?: readonly string[] } = {}) {
const granted = new Set(options.devicePermissions ?? []);
const hasActivePermission = vi.fn(async (_deviceId: string, permission: string) =>
granted.has(permission),
);
// Deliberately always true: if the evaluator ever consulted these for a device, every assertion
// below would flip, which is exactly what must not happen.
const isPlatformOwner = vi.fn(async () => true);
const hasActiveGrant = vi.fn(async () => true);
return {
evaluator: createAuthorizationEvaluator({
devices: { hasActivePermission } as never,
grants: { hasActiveGrant } as never,
platformOwners: { isPlatformOwner } as never,
}),
hasActiveGrant,
hasActivePermission,
isPlatformOwner,
};
}
describe("device principals go through the ordinary evaluator", () => {
it("allows a device holding the exact restricted permission", async () => {
const harness = evaluator({ devicePermissions: [displayPermission] });
await expect(
harness.evaluator.authorize({
baseline: "restricted",
permission: displayPermission,
principal: createDevicePrincipal(deviceId),
}),
).resolves.toEqual({ allowed: true });
expect(harness.hasActivePermission).toHaveBeenCalledWith(deviceId, displayPermission);
});
it("denies a device any permission it was not granted", async () => {
const harness = evaluator({ devicePermissions: [displayPermission] });
// Pairing a screen for one display is not authority over another module's.
await expect(
harness.evaluator.authorize({
baseline: "restricted",
permission: otherPermission,
principal: createDevicePrincipal(deviceId),
}),
).resolves.toEqual({ allowed: false });
});
it("never satisfies an authenticated baseline, however many permissions it holds", async () => {
const harness = evaluator({ devicePermissions: [displayPermission] });
// "A person is signed in" is not true of a television. Every route that settles for the
// `authenticated` baseline — the module UI inventory among them — stays closed.
await expect(
harness.evaluator.authorize({
baseline: "authenticated",
permission: displayPermission,
principal: createDevicePrincipal(deviceId),
}),
).resolves.toEqual({ allowed: false });
expect(harness.hasActivePermission).not.toHaveBeenCalled();
});
it("never consults platform ownership or user grants for a device", async () => {
const harness = evaluator();
await expect(
harness.evaluator.authorize({
baseline: "restricted",
permission: displayPermission,
principal: createDevicePrincipal(deviceId),
}),
).resolves.toEqual({ allowed: false });
// Both ports would have answered "yes". A device is denied anyway, because neither is a path a
// device can reach — there is no device that is a platform owner.
expect(harness.isPlatformOwner).not.toHaveBeenCalled();
expect(harness.hasActiveGrant).not.toHaveBeenCalled();
});
it("denies every device when the deployment composed no device support at all", async () => {
const withoutDevices = createAuthorizationEvaluator({
grants: { hasActiveGrant: async () => true } as never,
platformOwners: { isPlatformOwner: async () => true } as never,
});
await expect(
withoutDevices.authorize({
baseline: "restricted",
permission: displayPermission,
principal: createDevicePrincipal(deviceId),
}),
).resolves.toEqual({ allowed: false });
});
it("denies a scoped check for a device rather than treating it as unscoped", async () => {
const harness = evaluator({ devicePermissions: [displayPermission] });
await expect(
harness.evaluator.authorize({
baseline: "restricted",
permission: displayPermission,
principal: createDevicePrincipal(deviceId),
scope: "event-1",
}),
).resolves.toEqual({ allowed: false });
});
it("still allows a device through a public baseline, like any other principal", async () => {
const harness = evaluator();
await expect(
harness.evaluator.authorize({
baseline: "public",
permission: displayPermission,
principal: createDevicePrincipal(deviceId),
}),
).resolves.toEqual({ allowed: true });
});
it("leaves human and anonymous decisions exactly as they were", async () => {
const harness = evaluator();
await expect(
harness.evaluator.authorize({
baseline: "restricted",
permission: displayPermission,
principal: createAuthenticatedPrincipal("user-1"),
}),
).resolves.toEqual({ allowed: true });
await expect(
harness.evaluator.authorize({
baseline: "restricted",
permission: displayPermission,
principal: ANONYMOUS_PRINCIPAL,
}),
).resolves.toEqual({ allowed: false });
await expect(
harness.evaluator.authorize({
baseline: "authenticated",
permission: displayPermission,
principal: createAuthenticatedPrincipal("user-1"),
}),
).resolves.toEqual({ allowed: true });
});
it("propagates a device lookup failure rather than reporting it as denied", async () => {
const failing = createAuthorizationEvaluator({
devices: {
hasActivePermission: async () => {
throw new Error("database unavailable");
},
} as never,
grants: { hasActiveGrant: async () => false } as never,
platformOwners: { isPlatformOwner: async () => false } as never,
});
// A caller must never be told "no permission" when the true answer is "could not be
// determined", exactly as for a human.
await expect(
failing.authorize({
baseline: "restricted",
permission: displayPermission,
principal: createDevicePrincipal(deviceId),
}),
).rejects.toMatchObject({ code: "AUTHORIZATION_EVALUATION_FAILED" });
});
});