dotfiles VS git-secrets

From a technological point of view, TOR still has a couple of flaws which make it vulnerable to the metadata logging systems of ISPs:

- it needs a trailing non-zero buffer, randomized by the size of the payload, so that stream sizes and durations don't match

- it needs a request scattering feature, so that the requests for a specific website don't get proxied through the same nodes/paths

- it needs a failsafe browser engine, which doesn't give a flying damn about WebRTC and decides to actively drop features.

- it needs to stop monkey-patching out ("stubbing") the APIs that are compromising user privacy, and start removing those features.

I myself started a WebKit fork a while ago but eventually had to give up due to the sheer amount of work required to maintain such an engine project. I called it RetroKit [1], and I documented what kind of features in WebKit were already usable for tracking and had to be removed.

I'm sorry to be blunt here, but all that user privacy valueing electron bullshit that uses embedded chrome in the background doesn't cut it anymore. And neither does Firefox that literally goes rogue in an endless loop of requests when you block their tracking domains. The config settings in Firefox don't change shit anymore, and it will keep requesting the tracking domains. It does it also in Librefox and all the *wolf profile variants, just use a local eBPF firewall to verify. I added my non-complete opensnitch ruleset to my dotfiles for others to try out. [3]

If I would rewrite a browser engine today, I'd probably go for golang. But golang probably makes handling arbitrary network data a huge pain, so it's kinda useless for failsafe html5 parsing.

[1] https://github.com/tholian-network/retrokit

[2] (the browser using retrokit) https://github.com/tholian-network/stealth

[3] https://github.com/cookiengineer/dotfiles/tree/master/softwa...

Added lots of other helper methods, mostly for the shitty kind of CLI tools like yt-dlp, wget, tar, etc. My PS1 is a little more complex because I'm using emojis in the Terminal to represent states of repositories and to shorten the base paths. My complete bashrc is here if you're curious.

Personally I decided to have my dotfiles being not a git repo of the actual dotfiles, but more like a quick bootstrapping framework with the idea that the install process can be run repeatedly and incrementally without fucking already "patched" config files up.

Well, this doesn't really happen at places that don't suck. They had no least privilege access to critical secrets and no processes (like pre-commit hooks using git-secrets) to prevent them being committed.

# macOS brew install git-secrets # Linux git clone https://github.com/awslabs/git-secrets.git cd git-secrets make install

You could use Git hooks to prevent someone from being able to author a commit when you suspect there is a secret being committed. In addition to this, you could also perform this check server-side, in case someone did not run their Git hooks for whatever reason. For example, check out git-secrets.

git-secrets

Install git-secrets. Go into each of your repos, scan for past mistakes, and add a git-commit hook:

I agree that code scanning is really important, the best way to convince others is to identify high-risk threats in source code and present them to the decision-makers. For example, scanning Secrets is great for showing how repositories can be a massive vulnerability and identifying some low-hanging fruit, especially in the git history. Attackers are really after git repository access for this reason and there are plenty of open-source or free tools that you can use to illustrate the problem. Git-Secrets, Truffle Hog. These aren't great for a long-term commercial solution, something like GitGuardian is a better commercial tool but if the goal is just to illustrate the problem then finding some high-value secrets with free tools is a good way to convince the security personnel to invest in some solutions. Then the door is open to having more conversations as you have already proven the risk.

I worked for a big startup last year and was on a contract deadline for integrating a vendor framework into a React Native app.

It was taking too long to get a new temp demo license key and GitHub search with clever filters helped me track down a demo key that was recently uploaded to a test repo.

This is also why I use git-secrets in my repos.

https://github.com/awslabs/git-secrets

Under the covers, it is simply looking up an 'ignore' list stored in YML during each scan. If you are building your own, you might also want to see how AWS Labs is doing it in their solution git secrets.