Claude Security Public Beta - Opus 4.7 Validates Findings and Suggests Patches in Your Repo
0:00 / 0:53
News
Claude Security Public Beta - Opus 4.7 Validates Findings and Suggests Patches in Your Repo
calendar_today Date:
schedule Duration: 0:53
visibility Views: 103
database
Summary Report
Anthropic launched Claude Security in public beta for Enterprise customers, putting Opus 4.7 to work scanning codebases, validating findings, and proposing patches with no custom tooling needed.
- 01. Claude Security is now in public beta for Claude Enterprise customers, with Team and Max plan support coming.
- 02. It scans codebases, validates each finding adversarially to cut false positives, and proposes patches a human reviews and approves.
- 03. Hundreds of organisations used it during the February research preview, catching issues existing scanners missed.
- 04. Public beta adds scheduled scans, directory-level targeting, CSV and Markdown exports, webhook alerts to Slack or Jira, and persistent dismissals.
- 05. Anthropic uses Claude Security on its own codebase, and the product needs no API integration or custom agent build to run.
Anthropic has released Claude Security to public beta for Claude Enterprise customers, offering a code vulnerability scanning tool powered by Opus 4.7. The system activates through a simple toggle in the admin console, requiring no API integration or custom agent development.
Claude Security approaches code analysis like a human security researcher, tracing data flows through codebases and validating findings against adversarial counter-arguments to minimise false positives. When vulnerabilities are identified, the system proposes patches for human review and approval. Anthropic claims hundreds of organisations have tested the tool during its research preview phase since February, identifying issues that existing scanners missed.
The public beta introduces enterprise-focused features requested during the preview period. These include scheduled scanning capabilities, directory-level targeting for more precise analysis, and export options in CSV and Markdown formats. Integration features allow webhook alerts to be sent to Slack or Jira, whilst dismissal tracking ensures ignored findings don't reappear in subsequent scans.
Anthropic uses the same tool internally on its own codebase, lending credibility to the product's effectiveness. Whilst currently limited to Enterprise customers, support for Team and Max plan subscribers is planned for future releases. The company positions Claude Security as a solution to the noise problem that has historically plagued automated security scanning tools, betting that LLM-powered analysis can deliver more actionable results.
Meta Data
Company:
LLM:
Model:
