Pass a validator function to jwt-bouncer which will invoke the validator to check the jwt.
The Whitelist validator sample function takes an options object as its argument. The options object will contain a property named req which represents the ExpressJS request object and a url to the error documentation. The validator determines if a JWT is present on the incoming request as an Authorization header containing a Bearer token. If so, it decodes and checks the JWT against a whitelist.
This validator assumes that a whitelist array has already been retrieved by prior middleware and is available as a
whitelistproperty on the request object.
If everything is ok, the validator returns an object containing these properties:
ok- a value oftruedenotes successwhiteListItem- the found item in the whitelistdecodedToken- the decoded tokentoken- the token (not decoded)
{
ok: true,
whiteListItem: foundWhiteListItem,
decodedToken: decoded,
token
}If the jwt did not pass whitelist validation then the object returned from the validator will look like:
{
ok: false, err
}
The jwt validator sample function takes an options object as its argument. The options object will contain a property named req which represents the ExpressJS request object and a url to the error documentation.
The CDS Client (EHR Vendor) MAY make its JWK (JSON Web Key/Public Key) set available via a URL identified by the jku header field, as defined by rfc7515 4.1.2. If the jku (JSON Web Key URL) property does not exist on incoming JWT header from the client then attempt to obtain jku property from whitelist.
If the jku DOES exists then we fetch the JWK Set using the jku url. The required kid value from the JWT header allows a CDS Service to identify the correct JWK in the JWK Set.
If a jku DOES NOT exist, the CDS Client MUST communicate the JWK to the CDS Service provider (that's us) out-of-band. In this case, we should have already communicated with the EHR vendor and stored a single key from the EHR vendor as a string in PEM format using the jwkPublicKeyPEM property on the whitelist item.
Either way, once we have a JWK, we make sure its in PEM format and attempt to verify the jwt.
This validator assumes that a
whitelistValidatorResultproperty on the incoming request object.
If everything is ok, the validator returns an object with an ok prop set to true:
{
ok: true;
}If everything there's an error, then return an object with an ok prop set to false and an err prop:
{
ok: false,
err
}
If there's a problem, the validator should return a validation result object with an ok prop set to false and an err prop set to an error. The object is handled by the jwt-bouncer which is responsible for calling the error handling middleware.
{
"ok": false,
err
}
If there is not an error, the validator should return a validation result object with an ok prop set to true and any other properties that the next middleware in line may need, if necessary.
The jwt-bouncer will take the validation result object and set it to the designated property that was determined when you created the validator through the jwt-bouncer. Finally, the jwt-bouncer calls the next function and the next middleware in line continues to processes the request.