DeFi Intel

Visa deploys Anthropic’s Claude Mythos to hunt vulnerabilities

Visa's proactive use of AI for vulnerability detection sets a precedent for continuous adaptation in cybersecurity, enhancing global payment security.

TL;DR

Anthropic upgraded Claude Security on October 21, 2026 to scan code continuously with each commit and automatically track vulnerability fixes. The move tightens security coverage, reduces manual work, and includes credits for enterprise users, signaling a shift toward AI‑driven, workflow‑integrated security.

Key facts

  • Anthropic announced the Claude Security update on October 21, 2026.
  • The tool now rescans repositories automatically on every new commit.
  • Each vulnerability is tracked individually and marked fixed when the code change is detected.
  • Projects can include up to eight repositories from GitHub or GitLab.
  • Claude Enterprise customers can receive up to $25,000 in monthly usage credits for two months, valid until December 31, 2026.

Anthropic announced on October 21, 2026 that its Claude Security tool now continuously monitors code repositories, automatically rescanning with each new commit and tracking each vulnerability until it is resolved. The update, available in public beta for Claude Enterprise customers, expands the tool’s capabilities by using the full Claude model family, including Claude Mythos, and adds API and Model Context Protocol integrations. Eligible organizations can receive up to $25,000 in monthly usage credits for two months, with the offer running until the end of 2026.

Background

Claude Security, Anthropic’s AI‑driven code‑scanning product, entered public beta in April 2026 and quickly attracted thousands of organizations, which together reported more than one million vulnerabilities. The platform originally performed point‑in‑time scans on a scheduled basis, leaving gaps during which code could be shipped unchecked. User feedback highlighted the need for tighter, more automated coverage that fits existing development workflows. In response, Anthropic rebuilt the service to trigger scans on every repository commit, shifting from heavy baseline analyses to lighter, incremental checks after the initial deep scan. The tool also now tracks each finding individually, automatically marking a vulnerability as fixed when the corrective code is detected and reopening it if the issue reappears. Projects can link up to eight repositories from services like GitHub and GitLab, and findings can be exported via API or the Model Context Protocol, a standard for connecting AI models to external software. Alongside Claude Security, Anthropic is testing a Claude Code plugin for local, session‑specific scans and has launched an opt‑in open‑source scanner that has already surfaced 493 issues across 25 projects. These developments aim to keep security coverage current and reduce manual effort for development teams.

What happened

On October 21, 2026 Anthropic announced a major update to its Claude Security product, shifting from calendar‑based scans to continuous monitoring that triggers on each new commit, as reported by Cryptobriefing. The update introduces a baseline deep scan when a repository is first connected, followed by lighter incremental scans for subsequent changes, eliminating the need for manual scheduling. Each vulnerability discovered is now tracked on its own record; the system automatically marks a finding as fixed when it detects the corrective code change and can reopen the issue if the same flaw resurfaces later. Projects may include up to eight repositories from platforms such as GitHub and GitLab, and findings can flow into external tools through API and Model Context Protocol (MCP) connectivity, allowing teams to integrate results into existing workflows. The public beta remains limited to Claude Enterprise customers, who can receive up to $25,000 in monthly usage credits for two months, with the credit offer valid until December 31, 2026. In parallel, Anthropic is running a beta plugin called Claude Code that enables developers to perform local, session‑specific scans on particular repositories or individual changes. Since its initial public beta launch in April 2026, Claude Security has been used by thousands of organizations, identifying over one million vulnerabilities, while an opt‑in open‑source scanner has addressed 493 issues across 25 projects.

Market impact

The announcement includes a financial incentive of up to $25,000 in monthly usage credits for eligible Claude Enterprise customers, an offer that runs until the end of 2026. Since its public beta debut in April 2026, Claude Security has been adopted by thousands of organizations, collectively uncovering more than one million vulnerabilities. Anthropic’s open‑source scanner has already identified 493 issues across 25 projects, demonstrating the model’s reach beyond enterprise customers. The eight‑repository limit per project may influence how larger firms allocate their codebases, potentially prompting multiple project setups to achieve full coverage. API and MCP integration capabilities are positioned to drive adoption by allowing findings to be incorporated into existing security tooling, which could affect spending patterns on complementary security solutions.

Why it matters

Continuous, commit‑triggered scanning narrows the window during which untested code can be shipped, reducing the risk of vulnerabilities reaching production. Automatic resolution tracking cuts manual verification work, allowing security teams to focus on higher‑impact tasks. The ability to reopen issues if a fix regresses ensures ongoing protection against recurring flaws. By supporting up to eight repositories per project and offering API and MCP integrations, Claude Security aligns with existing development workflows, increasing the likelihood of adoption. The credit incentive lowers the barrier for enterprises to trial the technology before committing to larger spend, potentially accelerating the shift toward AI‑augmented security practices across the industry.

DeFi Intel’s take

Editorial analysis, not a statement of fact.

DeFi Intel's take: the shift to continuous, commit‑level scanning suggests Anthropic is positioning Claude Security as a core component of modern DevOps pipelines rather than a periodic audit tool. We read this as an effort to embed AI‑driven vulnerability detection directly into the software delivery lifecycle, which could reduce the time between discovery and remediation. The automatic fix verification and reopening features indicate a focus on minimizing manual overhead and preventing regression, a common pain point for security teams. The eight‑repository cap may prompt larger firms to segment their codebases, potentially creating new adoption pathways. Overall, the move hints at broader industry momentum toward integrating AI security tools with existing workflows, which could reshape how enterprises allocate budget for code‑security solutions.

Sources used for this analysis

DeFi Intel wrote this analysis from the reporting above. Any figure or quotation can be checked against those pages.

Frequently asked questions

What change did Anthropic make to Claude Security in October 2026?

Anthropic updated Claude Security to continuously monitor code repositories, triggering scans on every new commit instead of using a calendar schedule. The tool now runs a deep baseline scan when a repository is first connected and then performs lighter incremental scans for each subsequent change.

How does Claude Security handle vulnerability tracking after the update?

Each vulnerability is tracked on its own record. When the system detects the code change that fixes a flaw, it automatically marks the finding as resolved. If the same issue reappears later, Claude Security can reopen the finding, ensuring ongoing oversight.

What incentives are available for Claude Enterprise customers to try the new features?

Eligible Claude Enterprise customers can receive up to $25,000 in monthly usage credits for two months. The credit offer is available until December 31, 2026, providing a financial incentive to test the continuous monitoring capabilities.

Direct answers

Short answers, written to be quoted.

How does the new Claude Security handle code changes?
The updated Claude Security automatically rescans a repository whenever a new commit is made, performing lighter incremental scans after an initial deep baseline scan.
What happens when a vulnerability is fixed in the code?
Claude Security detects the corrective code change and automatically marks the associated vulnerability as resolved, reopening it only if the issue reappears later.
Are there any financial incentives for enterprises to use the new features?
Eligible Claude Enterprise customers can receive up to $25,000 in monthly usage credits for two months, with the offer running until the end of 2026.
Read full article on Crypto Briefing →

DeFi Intel is an entity-graph aggregator: we curate, tag and link crypto news to a typed knowledge graph of protocols, tokens, people and incidents. We do not republish the full article body. Use the link above to read the original report at Crypto Briefing.

Entities in this story

Discussion

Comments are reviewed before publication.

No comments yet.

Stay current on entity-tagged crypto news

Get the weekly DeFi Intel brief — entity-graph intelligence covering protocols, tokens and incidents, free to your inbox.