API Security Weekly: Issue #10

Vulnerabilities

Another API vulnerability has been found in Google+ (we reported on the previous one in our first newsletter back in October). Turns out that an update that Google rolled out in November put user data at risk because permissions were not properly enforced. The API could provide access to user profile data even if the data was not public. Google did a good job disclosing it: they found it themselves, fixed it in 6 days, and they report there’s no evidence of any exploit.

Unprotected APIs

Ethereum wallets and mining equipment are getting hacked through JSON-RPC API. The API was designed to be used by applications running locally on the same server, so it has no protection by default. Unfortunately, some systems using the API have all interfaces open, thus exposing the API without any security on port 8545. The API gives you full access, and there have already been reports of millions of dollars worth of the cryptocurrency stolen through the API.

Docker APIs are switched off by default, but they are required for remote administration and some 3rd-party tools. If the APIs are opened without applying proper security ,they leave your infrastructure vulnerable. Pierluigi Paganini has found more than 1000 exposed hosts, with plenty of them already compromised and running malicious workloads such as cryptomining.

Research

Researches from Texas A&M University used WARDroid framework to analyze the backend APIs of 10,000 popular Google Play apps. The researchers then used the APIs directly, bypassing the apps, and found 926 of the apps to be vulnerable to various API attacks. See the original research paper and its summary in ThreatPost.

Analysts

McAfee Labs have published their 2019 Threats Predictions Report. On API security, they have included API access to cloud-stored data in their predicitions: "Cloud-native attacks targeting weak APIs or ungoverned API endpoints to gain access to the data in SaaS as well as in PaaS and serverless workloads."

Best Practices

Aaron Lint from Arxan talks about how web application firewall (WAF) cannot detect API hacks when the attackers mask their behavior by hiding attack calls (like out of bounds parameters) in normal API traffic. This tactic will probably also fool other tools based on Artificial Intelligence/Machine Learning/anomaly detection.

You can subscribe to this newsletter at https://APISecurity.io.