Artificial Intelligence
Anthropic launches Cyber Mission for defenders
Anthropic said it was creating a Cyber Mission with a critical infrastructure programme and an open-source scanner, and said the effort would extend models, engineers and funding to outside defenders.
- Anthropic said it was launching the Anthropic Cyber Mission as a long-term effort to help secure software and systems.
- Its first two components were the Critical Infrastructure Defense Program and OSS Scanner.
- Anthropic said the infrastructure programme would bring Claude models, on-site engineers and threat research to partner organisations.
- Anthropic said OSS Scanner would send periodic model-generated reports to enrolled open-source projects without human review, and it said it expected more than 90% of those findings to be true positives.
- Anthropic said it had merged Project Glasswing into its expanded Cyber Verification Program and had already provided Claude support to more than half of US states.
Anthropic said it was starting the Anthropic Cyber Mission, which it described as a longer-term effort to help secure software and connected systems. It said the first two parts were a programme for critical infrastructure and a scanner for open-source projects.
For critical infrastructure, Anthropic said the new programme would bring Claude models, engineers and threat research to trusted providers that support operational technology. It named Accenture, Booz Allen, CrowdStrike, Deloitte, Dragos, Hitachi, Insane Cyber, Nozomi Networks, Palo Alto Networks, PwC and Rockwell Automation as founding partners.
For open-source software, Anthropic said OSS Scanner would offer enrolled projects regular scans from its strongest models at no cost. It said the reports would be model-generated and sent without human review, although it said that some would be inaccurate.
Anthropic said it had already combined Project Glasswing with its expanded Cyber Verification Program, and said it had offered Claude models and technical support to more than half of US states since June. It also said it had funded several open-source groups and would continue to expand the work to other areas.
Named in this story
Companies
- Anthropic
- announced the Cyber Mission and related programmes
- Accenture
- a founding partner of the critical infrastructure programme
- Booz Allen
- a founding partner of the critical infrastructure programme
- Dragos
- a founding partner of the critical infrastructure programme
- Nozomi Networks
- a founding partner of the critical infrastructure programme
- Palo Alto Networks
- a founding partner of the critical infrastructure programme
- Rockwell Automation
- a founding partner of the critical infrastructure programme
Organisations
- Python Software Foundation
- one of the open-source organisations Anthropic said it funded
- OpenSSF
- one of the open-source organisations Anthropic said it funded through the Linux Foundation
- Apache Software Foundation
- one of the open-source organisations Anthropic said it funded
- Akrites
- an organisation that collects and coordinates vulnerability reports
- Gold Eagle
- an organisation that collects and coordinates vulnerability reports
Products and systems
- Anthropic Cyber Mission
- umbrella effort to support defenders of software and systems
- Critical Infrastructure Defense Program
- new programme for infrastructure defenders
- OSS Scanner
- new service that scans open-source projects
- Project Glasswing
- earlier effort that found vulnerabilities and informed the new work
- Cyber Verification Program
- expanded programme that absorbed Project Glasswing
- Claude
- models used in the defence programmes
- Defender Advantage Fund
- fund launched in August to support pilots and keep OSS Scanner free
How the source tells it
The piece was promotional and forward-looking, with repeated claims about scale, capability and partnership alongside a strongly urgent risk narrative.
- vendor boosterism the company presented its own launch as a major long-term commitment and foregrounded its tools, funding and reach
- urgency and threat framing it emphasised exposed systems, state-backed adversaries and possible outages to make the risk feel immediate
- speculative futurism it offered a two-year forecast about AI favouring defence as if it were a likely trajectory
- collaborative reassurance it leaned on partnerships, trusted providers and support for defenders to present the effort as stabilising