Anthropic launches OSS Scanner, a free open-source vulnerability scanning service
11/04/2026 — 11/04, 20:20·1 sources (all)·1 reports (all)·In the 2026-11-05 briefing
Key takeaways
- Anthropic introduced OSS Scanner, a free opt-in service that uses its most capable model to regularly produce security reports for open-sour…
Background & analysis
According to a November 5, 2026 report by InfoQ, Anthropic has launched OSS Scanner, a free, opt-in service that scans open source projects for vulnerabilities. Maintainers who register receive periodic security reports generated by Anthropic's most capable models, including Claude Mythos. Reports are sent to maintainers without human review, and each project decides for itself which fixes to accept.
Citing an Anthropic announcement dated October 8, the report says the company's Frontier Red Team described human verification as the bottleneck. Over six months, its models flagged more than 29,000 potential vulnerabilities, but researchers manually reviewed and triaged only about 6,000. Maintainers who asked for early access have already received nearly 5,000 unverified reports together with the corresponding fixes. Every scan report contains a standalone reproduction program and an explanation of the suspected vulnerability, plus a binary-search tracing of the code change that introduced the defect where possible, and a candidate patch. The service is a continuation of the Glasswing program, which InfoQ covered in April 2026, when Anthropic gave Mythos preview access only to partner organizations.
Noah Misch, a PostgreSQL maintainer, said the fast-track access helped the project address the relevant issues before a formal release. Anthropic also tested an earlier version of the scanner by inviting penetration testers to evaluate 97 critical-severity findings across 48 projects. Of those, 85 met the coordinated vulnerability disclosure standard and 11 were genuine. The report adds that a high-severity curl vulnerability was discovered through the system and fixed in 8.23.0. So far, the available material consists of this single source, with no independent confirmation from Anthropic's own channels.
AI-generated from 1 reports · updated 4 hours ago
- Heat index
- 0
- All sources
- 1
- All reports
- 1
- First seen
- 5 hours ago
Reports on this story headlines open the original
Anthropic launched OSS Scanner, a free opt-in service that sends open-source maintainers unscreened security reports generated by its strongest models, each with a reproducer, a bisected culprit commit and a candidate patch. Anthropic says its models flagged more than 29,000 potential flaws in six months while researchers manually triaged only about 6,000; a curl HTTP/2 use-after-free found this way was fixed in curl 8.23.0.
InfoQ 中文AI score 64
Other stories people are talking about
- Heat index 48Irregular Security attributes AI agent attacks to its own misstep1 sources
- Heat index 563Anthropic nears RSI as AI compute obligations reach $895B14 sources
- Heat index 95NewGoogle Backs Anthropic-Linked Data Center Project with Credit Support1 sources
- Heat index 51OpenAI, Anthropic and others pledge to curb AI election disinformation1 sources
- Heat index 46Anthropic's 15-person chip team uses Claude in in-house chip design1 sources
How is heat calculated?About the methodHide
Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.
This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.
- Surge
- Discussion rising fast
- New
- First report within 6 hours
- Rising
- Still gathering discussion

