Since 2022, Google's Open Source Software Vulnerability Reward Program (OSS VRP) has been the path for outside researchers to get paid for reporting security flaws in the company's open source code, including projects like Flutter, Angular, Go, and Fuchsia.

With an announcement on X, they have now decided to discontinue the product-facing side of it.

They are calling the change temporary, while supply chain reports remain open and anything submitted before October 1 stays unaffected. What's closed, what's not?

Product vulnerabilities are bugs in the projects themselves, like a failing HTML sanitizer, memory corruption issues in file format parsers, or insecure code examples in documentation.

The updated rules do not limit the change to a specific project tier, as they just won't be accepting any reports related to this.

Supply chain reports, on the other hand, cover vulnerabilities in how the software is built and shipped. Exposed package manager credentials used to publish build artifacts is one case the rules page lists. It remains unaffected by this change.

There's also an exception for some Google Cloud repositories. If a bug there affects a Cloud product, Google may still accept the report, but through its Cloud VRP.

Why did it come to this?

Back in March, a post on the Bug Hunters blog from Google engineers said AI-generated reports were flooding the program. Some were serving up hallucinated information, while others were flagging legit coding errors that had little to no impact on the security posture of the targeted project.

Google's first response was to raise the bar on memory corruption reports for its two top project tiers. Researchers had to either reproduce the bug through an existing OSS-Fuzz fuzz target or point to a patch that maintainers had already merged.

An update to the same post the following month went further. The standard and low-priority tiers, OT2 and OT3, stopped offering rewards or credit for product vulnerabilities and other security issues. The top supply chain reward for OT2 projects also fell to $3,133.70.

Supply chain reports still pay, from $500 on OT2 projects up to $31,337 on flagship ones.

As an alternative, Google points to the Patch Rewards Program, which pays between $100 and $15,000, though only for patches that have stayed in a project for a month without being reverted.

An open question

Google has not said when or in what form product vulnerability reports will return. Their announcement only promises an update in the first quarter of 2027 while they continue reworking that part of the program.

There's also a loose end. At the time of writing, the OSS VRP page still lists reward ranges for product vulnerabilities, up to $7,500. Though, as you saw earlier, the rules page for it has already been updated, so it shouldn't be long before this is addressed.

no comments (yet)

sorted by: hot top controversial new old
there doesn't seem to be anything here
9 comments in c/technology@lemmy.world view that post
[–] 27 points 1 month ago (2 children)

Apparently Google doesn't care about good PR anymore.

https://en.wikipedia.org/wiki/Don%27t_be_evil

Don't Be Evil.
They have fully embraced it now.

  • source
  • hideshow 2 child comments
  • [–] 27 points 1 month ago

    Anyone finding a vulnerability now should really be submitting it to a broker instead of the company. Brokers pay way more for good exploits than the big guys (Google/MS/Amazon/etc) and won't just refuse to pay claiming you found an undocumented feature.

  • source
  • [–] 20 points 1 month ago (1 child)

    I had a hunch going into the article and I was right.

    I could have had some pity for those devs who are supporting the open source ethos being flooded with AI grifters spamming bug reports for the money. However that pity instantly goes right out the window given that they're working for one of the biggest companies pushing AI and why tech hardware is becoming more and more out of reach for the average Joe.

  • source
  • hideshow 1 child comment
  • I saw a PR to a project I follow that involves a specific piece if hardware. There are about 10 "solutions" in pr now of scammers trying to claim the bounty. None of them have even bothered to get the piece of hardware necessary to actually test their fixes. It's disgusting.

    I'm very happy for Claude starting to mark their code, and hope it becomes easier to detect. I'm actually all for ai being used to solve these issues, but I think it should make you ineligible to collect a bounty on it if you use AI.

  • source
  • parent
  • [–] 7 points 1 month ago (2 children)
  • 1 comment in c/opensource@lemmy.ml view that post
    [–] 4 points 1 month ago* (last edited 1 month ago)

    do they remember this exists because people were finding those bugs and exploiting them instead?

  • source
  • 1 comment in c/keralalinux@lemmy.ca view that post
    [–] 1 point 1 month ago

    This pause is due to a significant rise in automated submissions, the vast majority of which are not valid

  • source