The curl project is a project driven by volunteers with no financing at all except for a few sponsors who pay for the server hosting and for contributors to work on features and bug fixes on work hours. curl and libcurl are used widely by companies and commercial software so a fair amount of work is done by people during paid work hours.
This said, we don't have any money in the project. Nada. Zilch. We can't pay bug bounties or hire people to do specific things for us. We can only ask people or companies to volunteer things or services for us.
This is not a complaint - far from it. It works really well and we have a good stream of contributions, bugs reports and more. We are fortunate enough to make widely used software which gives our project a certain impact in the world.
Hacker One coordinates a bug bounty program for flaws that affects "the Internet", and based on previously paid out bounties, serious flaws in libcurl match that description and can be deemed worthy of bounties. For example, 3000 USD was paid for libcurl: URL request injection (the curl advisory for that flaw) and 1000 USD was paid for libcurl duphandle read out of bounds (the corresponding curl advisory).
I think more flaws in libcurl could've met the criteria, but I suspect more people than me haven't been aware of this possibility for bounties.
I was glad to find out that this bounty program pays out money for libcurl issues and I hope it will motivate people to take an extra look into the inner workings of libcurl and help us improve.
The bounty program is run and administered completely out of control or insight from the curl project itself and I must underscore that while libcurl issues can qualify, their emphasis is on fixing vulnerabilities in Internet software that have a potentially big impact.
To qualify for this bounty, vulnerabilities must meet the following criteria:
- Be implementation agnostic: the vulnerability is present in implementations from multiple vendors or a vendor with dominant market share. Do not send vulnerabilities that only impact a single website, product, or project.
- Be open source: finding manifests itself in at least one popular open source project.
In addition, vulnerabilities should meet most of the following criteria:
- Be widespread: vulnerability manifests itself across a wide range of products, or impacts a large number of end users.
- Have critical impact: vulnerability has extreme negative consequences for the general public.
- Be novel: vulnerability is new or unusual in an interesting way.
If your libcurl security flaw matches this, go ahead and submit your request for a bounty. If you're at a company using libcurl at scale, consider joining that program as a bounty sponsor!