update oauth_dispatch ADRs to accepted (#24474)

The following oauth_dispatch ADRs are being
marked as accepted, because they were implemented
and are in use in Open edX.
- 0006-enforce-scopes-in-LMS-APIs.rst
- 0007-include-organizations-in-tokens.rst

Given that, scopes have still not been widely
adopted in Open edX, which would require additional
decisions and investment.
This commit is contained in:
Robert Raposa
2020-07-14 09:57:30 -04:00
committed by GitHub
parent 3eb7b4e18d
commit e89e1a4038
2 changed files with 9 additions and 9 deletions

View File

@@ -4,7 +4,7 @@
Status
------
Proposed
Accepted
Context
-------

View File

@@ -4,7 +4,7 @@
Status
------
Proposed
Accepted
Context
-------
@@ -42,9 +42,9 @@ Here are a few organizational relationships that exist in the edX system:
do so either via
a. bulk APIs using the *Client Credentials grant type* (e.g., to
synchronize their own data in a background process) or
synchronize their own data in a background process) or
b. a user-specific API on behalf of a logged-in user via the
b. a user-specific API on behalf of a logged-in user via the
*Authorization grant type* and *edX as the identity provider*
(e.g., to display user-specific data on their own portal).
@@ -100,7 +100,7 @@ organization information to the granting end-user.
"content_org:Microsoft"
* For a token created on behalf of a user (*not* created via a
* For a token created on behalf of a user (*not* created via a
*Client Credentials grant type*), the token
is further restricted specifically for the granting user. And so, a
"user" filter with the value "me" would be added for this grant type.
@@ -137,7 +137,7 @@ Token Examples
Client Credentials (server-to-server) grant type
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
When a trusted application makes server-to-server calls, the application's
When a trusted application makes server-to-server calls, the application's
service user info is included in the JWT and the *filters* field
includes the organization identifier and type associated with the application.
@@ -166,7 +166,7 @@ filter “me” in the *filters* field.
"version": "1.0",
"preferred_username": "ajay_mehta",
...
}
}
Consequences
------------
@@ -191,7 +191,7 @@ Consequences
make sure there are no security issues introduced where old endpoints
that are not aware of the new filter do not enforce it. Possible
ways of doing so are:
* Endpoints that are highly security sensitive should reject any
token that includes an unrecognized filter.
@@ -206,7 +206,7 @@ Consequences
they don't recognize. For example:
"grades:read:content_org"
Additionally, this alternative would allow tokens to specify different filters
for different scopes.