Google Suspends Open-Source Bug Bounty Reports Amid Surge in Invalid Automated Submissions

Google has temporarily suspended the submission of product vulnerability reports under its open-source software bug bounty programme following a sharp increase in automated reports that the company says are largely invalid.
The suspension took effect on October 1, 2026, and affects security researchers seeking rewards for vulnerabilities discovered in Google’s open-source projects.
The company said the decision was prompted by a significant rise in automated submissions, with the overwhelming majority found to be invalid.
Google announced the temporary suspension in a post on X on October 1, but did not disclose how many automated reports had been received or identify the tools responsible for generating them.
The company also did not confirm whether artificial intelligence was responsible for the increase in submissions.
The affected programme is known as the Open Source Software Vulnerability Reward Program, or OSS VRP.
Under the programme, security researchers can report vulnerabilities affecting Google’s open-source software and, where eligible, receive financial rewards.
Google’s programme rules now carry a notice stating that product vulnerability submissions are temporarily suspended while the company reviews and restructures that part of the programme.
The company said it expects to provide an update during the first quarter of 2027.
However, Google has not announced when it will resume accepting product vulnerability reports through the programme.
The suspension does not affect reports submitted before October 1.
Supply-chain compromise reports also remain eligible for submission and rewards under the programme.
What Google Considers a Product Vulnerability
Google’s OSS VRP defines a product vulnerability as a design or implementation weakness in its open-source software that could significantly compromise the confidentiality or integrity of user data in software built using the affected code.
Examples include memory-corruption vulnerabilities in file-format parsers and path-traversal flaws.
The programme categorises covered projects according to four different levels based on their importance and sensitivity.
The two highest categories are known as flagship and important projects.
Before the suspension, product vulnerability rewards for flagship projects ranged from $500 to $7,500, while important projects had rewards ranging from $101 to $3,133.70.
Google has now removed those product vulnerability reward figures from its published programme rules.
The changes were reflected in Google’s public GitHub repository on September 30, one day before the company publicly announced the temporary suspension.
Google’s most recent repository classification lists 26 flagship projects and 47 important projects.
The flagship category includes major open-source projects such as Go, Angular, Flutter, Bazel and Protocol Buffers.
Although product vulnerability rewards have been suspended, researchers can still receive rewards for certain other categories of security problems.
Supply-Chain Vulnerabilities Remain Eligible
Google continues to accept reports involving supply-chain compromises.
Such vulnerabilities could allow attackers to interfere with a project’s source code, published packages or other components of the software distribution process.
Rewards remain available for these reports.
Other qualifying security problems, including exposed credentials that could provide write access, also remain covered by the programme.
The current reward structure includes payments of up to $31,337 for supply-chain compromises affecting flagship projects.
Important projects can attract rewards ranging from $1,337 to $13,337, while standard projects have rewards ranging from $500 to $3,133.70.
For other security issues, Google lists rewards of $1,000 for flagship projects and $500 for important projects.
Low-priority projects in the fourth category do not have listed rewards.
Where Researchers Can Report Vulnerabilities
Google has pointed security researchers towards other programmes and reporting channels while the OSS VRP suspension remains in place.
One option is Google’s Cloud Vulnerability Reward Program.
Certain vulnerabilities affecting open-source repositories maintained by Google Cloud may still qualify if they also affect Google Cloud products.
The Cloud VRP has its own rules for determining eligibility and reward levels.
Researchers can also make use of Google’s Patch Rewards Program.
Unlike the vulnerability reward programme, the Patch Rewards Program pays researchers for developing security patches rather than simply identifying vulnerabilities.
Rewards under that programme range from $100 to $15,000 for qualifying security patches.
For a patch to qualify, the relevant project maintainers must accept it, and the patch generally has to remain in the project for at least one month before it can be submitted for consideration.
Google has also advised security researchers to determine whether a vulnerability affects a product covered by one of its other reward programmes.
The OSS VRP rules specifically encourage researchers to consider the Cloud VRP and AI Vulnerability Reward Program when reporting security issues affecting projects closely connected to Google Cloud or artificial intelligence products.
However, Google’s suspension notice does not clearly state whether researchers can continue submitting product vulnerability reports to the OSS VRP without requesting a reward.
Individual Projects Have Their Own Security Channels
Some Google open-source projects maintain separate security reporting procedures.
The Go programming language, for example, accepts security reports directly through its security team.
Google’s broader GitHub organisation security policy also directs researchers to its vulnerability reporting channel.
The situation differs from project to project, meaning researchers are being encouraged to review the security policy of the individual software project before submitting a report.
Angular, for example, continues to identify itself as part of the OSS VRP and directs security researchers towards Google’s Bug Hunters platform.
As of October 6, its security policy does not identify another dedicated reporting channel.
Google Had Already Tightened Its Rules
The temporary suspension comes several months after Google introduced stricter requirements for some reports submitted through the OSS VRP.
In March 2026, the company began demanding stronger evidence for certain vulnerability submissions in an effort to reduce the number of low-quality reports.
One example of acceptable evidence was a security patch that had already been merged into the affected project.
The changes were introduced as Google faced growing concerns about researchers submitting reports without sufficient evidence that the claimed vulnerabilities could actually be exploited.
At the time, reports emerged that some submissions generated with artificial intelligence contained fabricated or inaccurate technical details.
This raised concerns about the increasing use of AI tools to automate vulnerability discovery and reporting.
While AI can assist security researchers in identifying genuine flaws, automated systems can also produce false positives and present theoretical weaknesses as confirmed vulnerabilities.
The resulting volume of low-quality submissions can place additional pressure on security teams that must manually examine each report.
Growing Concerns Over AI-Generated Security Reports
The latest suspension highlights a broader challenge facing bug bounty programmes as artificial intelligence becomes increasingly capable of analysing software code.
Large language models and other AI systems can rapidly examine large codebases and generate vulnerability reports.
However, researchers and security teams have warned that automated discovery does not necessarily mean automated accuracy.
A vulnerability report still needs to demonstrate that a security weakness actually exists and can have a meaningful impact.
Without human verification, automated systems can generate large numbers of speculative findings that consume the time of security teams without producing actionable results.
Google’s decision suggests that the volume of such reports has become significant enough to affect the operation of its open-source bug bounty programme.
Go Warns Researchers About Unfiltered AI Reports
Google’s Go project has also introduced specific guidance concerning reports generated with large language models.
The guidance, added to the project’s security policy in September, asks researchers not to submit AI-generated reports without first reviewing and filtering the results.
The Go security team acknowledged that large language models can successfully identify genuine security vulnerabilities.
However, the same systems can also produce convincing reports describing vulnerabilities that do not actually exist.
The project warned that researchers who submit large quantities of unfiltered AI-generated findings should not expect to receive credit for those discoveries.
The warning reflects a growing concern throughout the cybersecurity industry: AI can dramatically increase the speed at which security research is performed, but it can also increase the amount of noise that security teams must process.
What the Suspension Means for Security Researchers
Google’s temporary suspension does not mean that security research on its open-source projects has stopped.
Researchers can still identify vulnerabilities and report eligible security issues through other Google programmes or project-specific channels.
Supply-chain vulnerabilities remain eligible under the OSS VRP, while researchers can also explore Google’s Cloud VRP, AI VRP and Patch Rewards Program where applicable.
The major change is that product vulnerability reports affecting Google’s open-source projects can no longer currently be submitted through the OSS VRP for rewards.
Google has indicated that it plans to provide another update in the first quarter of 2027.
Until then, researchers will need to pay close attention to the rules governing individual Google projects and other vulnerability reward programmes.
AI Is Changing the Bug Bounty Landscape
The development is another indication that artificial intelligence is changing the traditional relationship between security researchers and bug bounty programmes.
For years, automated security tools have helped researchers scan applications, identify weaknesses and generate technical reports.
The latest generation of AI systems has significantly expanded those capabilities by allowing models to analyse code, reason about potential attack paths and produce detailed vulnerability reports.
But the technology also introduces a new problem.
When thousands of researchers can use AI to generate vulnerability submissions at unprecedented speed, bug bounty programmes can become overwhelmed by findings that have not been properly verified.
Security teams must then spend additional resources determining which reports represent genuine vulnerabilities and which are false positives.
Google’s decision to temporarily pause product vulnerability rewards for its open-source projects appears to reflect that growing challenge.
The company has not said exactly how it will redesign the programme, but the planned update in early 2027 could provide more information about how automated and AI-assisted vulnerability research will be handled.
For now, the message to security researchers is straightforward: identifying a potential vulnerability is only the beginning. Reports must still be carefully validated, technically accurate and supported by credible evidence before they can provide meaningful value to developers and security teams.
As AI becomes more deeply embedded in cybersecurity research, the ability to separate genuine vulnerabilities from automated noise is likely to become increasingly important for both bug hunters and the organisations running bounty programmes.



