Know what you are connecting.

What a scan processes

Sentinel analyzes supported infrastructure configuration files that you upload or select through a repository connection. Its rules inspect configuration for security, reliability, and maintainability issues. Findings include the affected file or resource, severity, an explanation, and a recommendation.

In the hosted workflow, submitted file paths and contents are stored with the scan so that reports and supported remediation previews can be retrieved. Findings, scores, and scan history are associated with the account’s organization. Some browser history also remains on the device you use. Treat configuration files as data you are sharing with the service; remove credentials and unrelated sensitive material before uploading.

Repository access depends on the action

Reading a repository lets Sentinel retrieve configuration for analysis. Creating a fix branch and pull request requires write permissions. The permissions shown by your Git provider depend on the connection method; a connection that supports pull requests should not be treated as read-only access.

Review the authorization screen and the repositories available to the integration. Prefer access limited to the repositories you intend to review wherever your provider supports it. You can revoke a connection through your Git provider’s application or token settings.

Supported fixes can be previewed before a pull request is created. Review the diff, run your checks, and use your normal approval process before merging. If you enable automatic pull requests for a repository, review that setting as part of your team’s change policy.

Rules first, optional AI context

The core findings come from deterministic rules. Optional AI explanations use the configured AI provider to interpret finding details, including rule identifiers, descriptions, recommendations, and file or resource locations. Finding chat also sends the conversation and relevant finding and scan context to that provider.

Those details can contain information about your infrastructure. Do not paste passwords, tokens, or confidential material into chat. AI availability depends on the plan and service configuration; AI text remains guidance for an engineer to review.

Access and the limits of a report

Account authentication and organization access checks protect the hosted scan workflow. A configuration review describes the submitted files at the time of the scan. It does not establish the effective permissions of a running cloud account, prove that a deployment matches the repository, or guarantee that an application is secure.

Framework references in a report can help organize review work. They are not a certification or an independent audit. Validate changes in your environment, especially when a recommendation affects networking, storage, process privileges, or availability.

Data questions and security reports

Read the privacy policy for the service’s personal-data terms and contact details. For questions about storage, access, deletion, or a suspected security issue, email sergey@nerden.su. Include the affected feature and a safe description of the issue; do not send live secrets or customer configuration in an initial report.

Before adopting Sentinel for a client or regulated workload, confirm that the available permissions and processing terms fit your requirements. The infrastructure review guides explain practical checks you can pair with a scan.