node-token-introspection

Node token introspection package introspects a token towards an oauth service that follows the RFC 7662.

Install

npm install token-introspection --save

Node version

Currently we only support latest Node LTS. If you want to use an earlier version of node, please use babel register.

Usage

Introspect package is configured with endpoint and client credentials, and a function is returned. Calling that function with token, and optional token_type_hint will return a Promise.

const tokenIntrospection = require('token-introspection')({
    endpoint: 'https://example.com/introspect',
    client_id: '<Client ID>',
    client_secret: '<Client Secret>',
});

tokenIntrospection(token).then(console.log).catch(console.warn);

Configuration

| Field | Required | Comment | | ------------------------- | :------: | ------- | | jwks | (X) | Static JWKS of trusted keys, for example {keys: [{kty:'RSA',n:'4-4mhUVhY2k',e:'AQAB'}]} | | jwks_uri | (X) | URL of a trusted JWKS, for example https://example.com/jwks | | endpoint | (X) | URL to call, for instance https://example.com/introspect | | allowed_algs | | List of allowed signing algorithms, defaults to ['RS256'] | | jwks_cache_enabled | | If jwks response should be cached, defaults to true | | jwks_cache_maxentries | | How many jwk's to cache, defaults to 10 | | jwks_cache_time | | How long a jwk is cached, in ms, defaults to 5 min | | jwks_timeout | | Timeout in ms for fetching jwks, defaults to 10s | | jwks_ratelimit_enabled | | If ratelimit of calls to jwks endpoint, defaults to true | | jwks_ratelimit_per_minute | | Limits of jwks calls, defaults to 60 rpm | | client_id | | Client ID used to introspect | | client_secret | | Client secret used to introspect | | access_token | | Access token used to introspect, instead of client credentials | | user_agent | | Defaults to token-introspection | | fetch | | Defaults to node-fetch, but you can inject zipkin-instrumentation-fetch. |

At least one of the required configuration parameters jwks, jwks_uri or endpoint must be specified.

Flexibility in fetch

As you can provide your own fetch implementation, it is possible override the agent fetch uses for various purposes. These purpose can be things like zipkin/tracing, self signed certificates, client TLS authentication, proxy, adding a keepAlive, etc.

const HttpsProxy = require('https-proxy-agent');
const proxy = new HttpsProxy(proxySettings);

const customFetch = (endpoint, options) => {
    options.agent = proxy;
    process.env.HTTPS_PROXY = proxy;
    return fetch(endpoint, options);
};

const tokenIntrospection = require('token-introspection');
const introspector = tokenIntrospection({endpoint, ..., fetch: customFetch});

Errors

This is a promise/async library, and will resolve with success or reject with an Error subclass.

• IntrospectionError: Base error, thrown when introspection fails for some reason.
• ConfigurationError: Thrown when configuration is wrong.
• MalformedTokenError: Thrown when token is malformed, currently not publicly exposed.
• TokenNotActiveError: Thrown when token is not active, base error for TokenExpiredError and NotBeforeError.
• TokenExpiredError: Thrown in local introspection when token has expired.
• NotBeforeError: Thrown in local introspection when token is not yet valid

Showing debug output

Set the environment variable DEBUG=token-introspection.