Ring middleware for parsing, decoding and verifying a JWS-signed JWT token from the incoming request.
NOTE: There is no explicit support for JWE currently.
Built on top of the excellent auth0 JWT library.
Once wired into your ring server, the middleware will:
Authorization
HTTP header but this behaviour can be overridden using the find-token-fn
setting (see usage below).:claims
key on the incoming request.:claims
map to the request if no token is found.401
if the JWS signature in the token cannot be verified.401
if the token has expired (i.e. the exp claim indicates a time
in the past).
leeway-seconds
setting (see usage below).401
if the token will only be active in the future (i.e. the nbf claim indicates
a time in the future)
exp
, the leeway-seconds
setting can be used to introduce a leeway on this check.IMPORTANT: by default a missing token will simply cause authentication to be skipped - to override this behaviour set the
reject-missing-token?
setting (see below). This to allow for the common scenario where authentication is only required on a subset of endpoints (e.g. to skip authentication on a healthcheck endpoint). It is therefore expected that any token would then be used for authorization: and a missing token would cause this authorization to fail.
(require '[ring.middleware.jwt :as jwt])
(defn handler [request]
(response {:foo "bar"}))
(jwt/wrap-jwt handler {:issuers {"https://some/issuer" {:alg :HS256
:secret "asecret"}
"https://another/issuer" {:alg :RS256
:jwk-endpoint "https://some/jwks/endpoint"}
:no-issuer {:alg :HS256
:secret "anothersecret"}}})
Options:
:issuers
(mandatory): A map of issuer->cryptographic algorithm configuration. When receiving a JWT token, the middleware
will pull the issuer from the iss
claim and use it to lookup the appropriate algorithm in the middleware configuration to verify
the JWT. (So, the iss
claim is implicitly only "trusted" if verification succeeds.)
:no-issuer
key in the map of issuers - this value will be used if no iss
claim is found
in the incoming token.:find-token-fn
(optional): A single-argument function that will be used to pull the (encoded) token from the request map. If unspecified
the token will be sought from the bearer token given in the Authorization
header (i.e. an Authorization
HTTP header of the form "Bearer TOKEN"):reject-missing-token?
(optional): A flag indicating whether a request missing a JWT token will be rejected with a 401
response. Default is false
-
i.e. by default a missing token will simply cause authentication to be skipped.Depending upon the cryptographic algorithm, a different map of options will be required. Note that, at the point your ring middleware is wired up, ring-jwt will throw an error if it detects that the given options are invalid.
Currently the following JWA algorithms are supported for the purposes of JWS:
Algorithm | Options |
---|---|
ECDSA using P-256 and SHA-256 | {:alg :ES256 :public-key public-key} [1] |
RSASSA-PKCS-v1_5 using SHA-256 | {:alg :RS256 :public-key public-key} [1] |
{:alg :RS256 :jwk-endpoint "https://your/jwk/endpoint"} [2] | |
HMAC using SHA-256 | {:alg :HS256 :secret "your-secret"} |
public-key
is of type java.security.PublicKey
.Additionally, the following options are supported for all issuers:
leeway-seconds
: The number of seconds leeway to give when verifying the expiry/active from claims
of the token (i.e. the exp
and nbf
claims).Keys for use with Integrant or Duct are available in kelveden/duct.middleware.ring-jwt.
Copyright © 2021 Ovo Energy Ltd.
Distributed under the Eclipse Public License, the same as Clojure.
Can you improve this documentation? These fine people already did:
Alistair Dutton, Oliver Boyle, Jacek Schæ & Newton BeckEdit on GitHub
cljdoc is a website building & hosting documentation for Clojure/Script libraries
× close