Please don't commit .env

Some Dood on December 19, 2018

Let's face it. .env files are amazing. They have an easy-to-read syntax that stores all of our essential configurations into one file. Not only t... [Read Full]
markdown guide
 
  1. Install the operating system
  2. edit your user global gitignore vim ~/.gitignore
  3. paste:
.idea/
.vs/
.env
.pk
.pem
.pub

Do not think you will not make a mistake, be proactive and presume you will, we are only humans.

Edit: this will prevent adding to your project your IDE config files, environment and private keys

 

Itβ€˜s a good idea to put such things into your global .gitignore.

But keep in mind that you mostly work in projects with other people, so you have to put those things into your repositories .gitignore too, to avoid that your co-workers commit secrets.

 

That's pretty dangerous advice because it depends on how you configure your application.

I personally commit .env files because they don't contain sensitive information and typically having settings important to being able to start up my app in development. I have a feeling just about anyone using Docker Compose would also commit .env files too since it's used for setting your project name.

I reserve .env.prod or other environment specific files for sensitive API keys and other things I wouldn't want to commit.

 

It is not dangerous and also .env.prod is very specific.

You can always commit the files from gitignore, but you have to do it explicitly, so you avoid mistakes --force.

I just mean putting it into your main .gitignore file is dangerous because you run the risk of not committing essential settings to make your project work if it depends on .env being around.

Using --force every time they want to add a file isn't a viable solution IMO. That's a command you might run 50 times a day.

In other words, you should add .env on a per project basis when your project may have .env files that contain sensitive info, just like you would want to ignore any other files that have sensitive data.

 
 

Nope, thats what they say but is it a bad practice.

Each env had its own tweaks and sometimes diffenrent platforms, tasks, settings and folders. Each dev should have the liberty of using its own IDE in their own productive way.

Ah yes, let's not use the tool as instructed. That will go great.

Each env had its own tweaks and sometimes diffenrent platforms, tasks, settings and folders. Each dev should have the liberty of using its own IDE in their own productive way.

Yeah, and they tell you exactly which things to exclude in order to get that.

You should include the idea folder. End of story.

No, sorry.

Do some research on the topic you will see that many other developers share my opinion. Do not blindly trust the IDE creators, for them is a business, for you is just a tool.

IDE's are local tools, they are not related to the project. A big project then will have

.vim
.idea
.eclipse
.atom
.vscode

folders. Where did you saw that?

You can also browse popular open source projects source code and see that most of them do not have IDE config files.

 
 

Best and most useful solution. prevention before cure.

 

What's more alarming about this is that there are still some others who have not reverted their commits.

Just to clarify for anyone reading, a git revert will not solve the problem: you'll still have the .env file visible in your public git history. If you've accidentally committed a secret, you need to remove that commit completely.

If you catch the problem immediately, you can do a git reset HEAD^ followed by git push -f. This deletes the most recent commit and overwrites the remote Git history. If you realize too late what you've done, you'll need a slightly more complicated method (see this useful blog post for pointers).

Of course, on the Internet this is all pretty optimistic; someone could have scraped your secrets while they were publicly visible, even if only for a handful of seconds. The only way to really get your security back is to go change all your secret keys and passwords.

 

Not to mention that git push -f opens the door to all kinds of accidents.

 

Yup! That's true. If at any time you feel that you have been compromised, never hesitate to reset those keys and passwords. In fact, it should be the first thing you should do before messing up the commit history. The less time the API key is valid, the better it will be for everyone (except potential hijackers).

 

There's sooooooo... many unaware user be out there , anybody can literally bring their whole business to an end! πŸ˜“

I tried to talk with some of them and make them aware, but got no response for months..πŸ˜‘

Seems like they're like,

What! I don't care about a small .env file, I'm Rich, I don't care!πŸ˜‘

 

You could always show them a demo of the amount of damage you can do if you had access to that "small" .env file. I'm sure they'll be alarmed after that.

 

Will Do.

I've seen many juicy leaks there such as Google Cloud Platform Keys , Facebook Credentials , GitHub Credentials and even someone's phone number πŸ˜‚

 

Be very careful when doing this as some wildly misguided company may accuse you of hacking and try and prosecute.

 

A (foolish) friend of mine once hardcoded my Mailgun keys in the project code and pushed it to a public GitHub repo. Well, guess what, I had a $600 bill, with about 1.6 million emails sent over the next few hours. When I got to know this, I was dumbstruck and contacted Mailgun, who said that my account has been compromised by a spammer and even gave me the link to the code that contained my keys!

However, they insisted that I clear the $600, since it was usage on my account anyway. I didn't, but they kept sending me reminders.

Moral of the story: Beware of Mailgun customer "support".

 

As if it was fault of Mailgun... It's your friends fault and yours for giving him these credentials.

 

As if it was fault of Mailgun

I never said that.

It's your friends fault and yours for giving him these credentials.

We were doing a project together and I was the one that created the Mailgun account. How could I have known in advance he'd just hard-code it into the source files?

The point is that these issues can be handled more sensibly by a company. AWS does this all the time, writing off bills caused by DDoS attacks eating up all the bandwidth, etc. Mailgun were the ones who found out that the account was compromised, told me that it was used by a spammer, and then insisted that I pay the bills. Great experience for me!

 

Oh, no. That's just really unfortunate. Do they still remind you to this day or did they terminate the service?

 

They followed up for a really long time, sent several emails, and even tried to charge my credit card (which I had blocked by then). As of now, they've disabled my account. πŸ€“

I had half a mind to write a blog post about this incident and tweet it, but then decided to let it go.

Wow. It makes me wonder what the hijacker did to rack up a bill of $600.

They sent emails like "you've won a $1 million lottery" and other such scams. Somehow, I'm sure that of all the 1.6 million that were affected by this campaign, I lost the most!

How typical of them. Hays... πŸ€¦β€β™‚οΈ

 

I see no point in the general advice, not to commit a .env-file per se: I do it frequently and wondered what this post is about.
It is not about committing a standard configuration in your repo, which could be overriden e.g. by a .env.local. It is about not storing secrets in your repository. That's another topic and indeed best practice.

So my usage of .env seems to differ: providing a sane default configuration, which can and should be overridden and does not contain any secrets whatsoever.

For managing secrets in a repo there is an interesting approach by stackoverflow: github.com/StackExchange/blackbox
which unfortunately I had not the time trying, but looks interesting.

 

Yes, it's definitely okay to commit the .env file if and only if they contain general and not-so-sensitive information. Otherwise, the point of this article is to remind people that they should think twice before they commit .env files because of the serious ramifications that come with it.

 

The only thing I don't like about doing this is that when I change the defaults, they now show up as a diff. On some of my smaller projects I SSH onto the server and just pull the latest changes, so if I've made a change to my default .env file I have to worry about conflicts. I can avoid this problem in production by having the defaults be my production settings, but then I constantly have a dirty file in development.

The best solution to this in my opinion is to not commit .env and instead commit a .env.example. Then all you have to do is cp .env.example .env and you have the exact same things as before but without a dirty file when you change settings.

 

Hmmm. Interesting. I came into first contact with .env files via vuejs. It seems there, that you have a .env file which acts like your ".env.example" and could have a .env.local which contains overrides.

But doing some research that seems an uncommon practice πŸ€”

Mostly there is one .env and it isn't committed.

 

Why have it in your git directory at all? Most frameworks let you override the config location so set it to ~/.app-name/.env or similar. If you need to provide an example dummy config you can check in an .env.sample file.

 

Yes, that's true. At least for me, I think it's just easier to have a .env file because it requires minimal setup and messing around with global configurations.

At least in Node.js, all you have to do is npm install and require the dotenv package. Then in your code, just invoke the dotenv.config() and it should all be running smoothly via the process.env object. This way just saves you from the little extra effort you have to do with the nitty-gritty configurations.

But to each its own. Whatever workflow works the best for you, you should apply it, not because everyone does it but because you feel productive with it.

 

Take a look at a package I wrote on NPM: (envup). It allows you to version control the structure of your environment as a separate file making it easy for others to setup their own .env file without commiting any of the data to git.

 

I noted this a few years ago, before ".env" became the norm, but the effect was the same. People were putting their OAuth keys into configuration files, committing them to git, and a bot search came across several thousand exposed keys just looking for Amazon's. Similar numbers probably would have been found with google and others.

In that case, I wanted to make sure that my users knew where to put their stuff, so I created a credentials.template file that showed the format, and that got committed, but my own credentials did not. One could do the same here by having a README.env.txt file to document what to do, and cat that file to the console in an npm post-install hook.

The negative of that, though necessary, is it means you're not distributing running code. They can't just pull your files down and npm start and everything works. They have to finish the init by creating their own files. It may also complicate automated testing systems that would have to be configured to provide that file before running.

If you have made this mistake already, one possible way to fix it is to interactive rebase back to the sha that introduced the problem, wipe the file and add the .gitignore line there, and then deal with the merge conflicts as it pushes the rest up if you ever had to touch that file again (either in format or in updating the data in it).

Of course, how much work that is depends on the age of your code (how many commits and how many branches).

 

I agree 100% in any real project.

That said I have a repo on github right now, symfony4-impressions, that I explicitly intended to commit the .env and give to people. In this case I wanted people to have the foobar password for the docker mysql container so they can easily run the performance benchmarks to see for themselves if they wanted.

 
 

That's good and all, but as @matthias mentioned, it's still important to add them to your repositories because other people will also work on them. They might not have configured their global .gitignore like you have, which might cause them to accidentally commit the .env file.

Itβ€˜s a good idea to put such things into your global .gitignore.

But keep in mind that you mostly work in projects with other people, so you have to put those things into your repositories .gitignore too, to avoid that your co-workers commit secrets.

 
 

Oh, and this doesn't just apply to auth and application keys from cloud providers. It can also apply to the credentials you use for demo and testing servers for your client-side app, and database credentials for server-side apps.

 

I faced the same issue because of our intern. He published .env file for everyone on the internet.

The .env file contains mail, google cloud and bank API keys. It was really terrible.

 

How did you guys manage the situation? Did everyone scramble to reset the API keys while others scrambled to clean up the working tree and commit history?

 

I regenerate the mail and bank API keys. I also bank added IP limit to bank API portal.

Our lead was so angry. I created a script to remove critical commit histories before we faced this situation (about 2 years ago). I ran that shell script.

But it was so dangerous. Normally we don't have published repositories.

I can say this was our fault.

I take it that the intern didn't stay there for long...

 

And you can also use dotenv-linter to lint your .env files for consistency and best practices: github.com/wemake-services/dotenv-...

 

If .env files are so popular, git should warn before such file is commited. Writing a validation could help millions of developers

 

Now that you mentioned it, I'm surprised that it hasn't been implemented yet.

 

.env is an anti-pattern to me, because it requires overhead to keep it secure.

 

It seems to be quite a popular anti-pattern nowadays. πŸ˜…

Besides, adding one line to the .gitignore file shouldn't be that much of an overhead.

 

110% this. Thank you for sharing Some Dood. And nice GitHub search results to boot! :)

 

Thank you! I appreciate it. On the note of the GitHub search results, I was really disturbed by how many commits are being made with a similar title. I just felt the need to write about it as a PSA and a friendly reminder to all that they shouldn't commit the .env file unless it's there for a justifiable reason, such as the case with @buphmin .

I agree 100% in any real project.

That said I have a repo on github right now, symfony4-impressions, that I explicitly intended to commit the .env and give to people. In this case I wanted people to have the foobar password for the docker mysql container so they can easily run the performance benchmarks to see for themselves if they wanted.

 

Thank you for the tip on how to get my projects to the top of the search results...

 
 

Yes, it could be okay to commit there. At least it's not public. As long as you really trust the cloud service hosting your repository, then it should be fine. No guarantees, though. It's a risk you take when you put something in the internet.

 
 

This is really intricate and interesting. Thanks for sharing! It was a great read (yet very disturbing 😬 from a security standpoint).

 
 

I agree with the post. It really comes down to resetting and revoking the validity of compromised API keys. Never mind messing up with the commit history; just leave it as is and reset those keys ASAP.

 

I looked into this for a project a while ago. I think it's an interesting solution for securely sharing secrets among people who are authorized to configure production servers.

But I also believe that those secrets shouldn't be shared across the entire team to avoid disaster mentioned in some of the anecdotes in this thread, specifically the one about the intern publishing production keys in an .env file. That intern and other devs shouldn't have ever access to those keys in the first place.

I feel that anyone who needs to have access to 3rd party API credentials for local development should just try to generate their own, by signing up developer accounts with those 3rd party web services and registering their own API clients.

It's overhead. But it only needs to be done once and makes absolutely sure that a leak of such credentials is not going to cause any major issues.

In case there is no easy way to sign up yourself (e.g. payment processor APIs) an encrypted store via Git like that makes sense for sharing sandbox environment credentials.

Ideally I believe that .env files and similar configuration files that contain confidential production credentials should be stored in the (encrypted) storage of the CI server and only bundled on build & deploy. A single source of truth, and a single possible point of failure.

code of conduct - report abuse