Tools AI-Found Vulnerability Register aifv-0067
Register entry
Anthropic updates Glasswing’s aggregate vulnerability findings
- Project
- Glasswing partner and Anthropic open-source investigations
- Finder
- Claude Mythos models in partner and Anthropic investigations
- Autonomy
- Vendor-reported partner investigations and Anthropic open-source scanning; workflows vary
- Evidence tier
- Aggregate claim
- Bounty
- None published
- Confidence
- SINGLE
What happened
Anthropic reports at least 129,000 verified findings from partners between April and July, plus 5,500 from its own open-source work between April and October. More than 33,000 across these findings were rated high or critical.
Credit as recorded: Anthropic Cyber Verification Programme announcement, 6 October 2026. The partner figures come from a partial survey of 33 organisations using different triage approaches. Patch reporting is incomplete. This is an aggregate vendor claim, shown separately from identified CVEs and the coordinated-disclosure dashboard. Date basis: vendor aggregate update.
Sources
- Anthropic programme announcement and findingswww.anthropic.com/news/cyber-verification-program
One record from the AI-Found Vulnerability Register, maintained by the Security Desk.Data: CSV · JSON ·RSS · CC BY 4.0 with attribution to YFarmX.Tracker updated · 8 October 2026