I am sure that you are doing many things to secure your cloud applications and your public endpoint.
Are you doing the same things for the code that you push to a repo?
In this article, we discover what are the tools and what are the actions that can be done to ensure that no password, no access token or string configuration reach the repository.
Scenario
Imagine that you are a developer working for a customer application hosted inside Azure. The new feature that you are working on includes premium media content available only for users who pay the subscription.
You design the application so that premium users get access to the content using a Shared Access Signature. Meanwhile, you keep the storage account key for production environments in a separate configuration pipeline script. The account key is used to generate a Shared Access Signature Token. Once the configuration pipeline script is working, you push to the public repository.
Boom! in a few seconds, the storage account key is cached by a Gittyleaks bot, and your premium media content is copied and shared over torrents.
Context
The same thing can happen with any other secrets. Even if you are using the Azure Key Vault or App Configuration, you need to be careful where you store secrets and if and how you push them to a repository.
Even if it is a private repository, I would treat it as a public repo and ensure that no secrets are stored inside it. The security breach can be caused by another team member, which computer is compromised, and the attacker starts to search all the repos where the developer has access.
What can we do?
Many tools on the market can scan your repo or your commits before a push to ensure that no secrets are stored or pushed to it.
Some actions can be done to ensure that no secrets are stored inside a source control repository:
- Configure a tool to actively scan your repositories all the time for files that contain secrets
- Configure on each machine that push content to the repository a tool that scans the commit and deny the ones that contain secrets
- Integrate the tool in the pipelines
There so many other things related to WAF (Well Architecture Framework) and Cloud Governance that can be done (e.g. service principles and RBAC)
Tools
One of the tools that I like to use is git-secret that enables us to scan for secrets for AWS, Azure, and GCP. The open-source projects scan for passwords and any other sensitive information inside a repo. During the push process, the scan runs, preventing us from pushing sensitive information to repositories like GitHub, Bitbucket, or even TFS if you really want it.
Dev Machine
A development machine shall be configured in such a way to scan commits before a push for secrets.
(1) Place git-secrets somewhere in the PATH to be easily accessible by git
(2) ./install.ps1 | Command to install git-secrets on a Windows machine
(3) cd /path/RaduVRepo/IoTHome | Navigate to the repo that you want to protect. You need to do this action for each repository that you want to secure
(4) git secrets install | Install the tool
(5) git secrets -register-azure | Register the Azure plugin
(6) git secrets -register-aws | Register the AWS plugin
(7) git secrets - register-gcp | Register the GCP plugin
DONE!
The plugging is scanning the commits and notify us if we have secrets inside the code.
Active Scan
We can use 'git secrets --scan repoPath' to scan a specific repository for secrets. The command can be easily integrated inside a pipeline. Another approach is to scan every night all the repos to ensure that there are no secrets push to the repos.
Pipeline Integration
In pipelines, one of the step is to run git-secrets and ensure that things are clean. If not, remove the secrets and even remove the commit :-)
Final thoughts
Managing secrets, especially cloud secrets it's not an easy job. It is mandatory to use tools like git-secrets to scan the commits for secrets and integrated them inside your pipeline.
Comments
Post a Comment