ScannerBrightly for developers

Connect your assistant to verified-domain security findings, source and dependency scans, and CVE alerts.

Model Context Protocol

Connect ScannerBrightly to Cursor, Claude, or ChatGPT

Use the hosted Streamable HTTP server at https://mcp.scannerbrightly.com/mcp. Your MCP client completes Authrice OAuth and grants access to one ScannerBrightly organization. API keys are separate and are not used by the MCP connector.

{
  "mcpServers": {
    "scannerbrightly": {
      "type": "http",
      "url": "https://mcp.scannerbrightly.com/mcp"
    }
  }
}

OAuth scopes and authorization

Read tools require scannerbrightly:read. Starting scans, acknowledging alerts, and creating alert tickets require scannerbrightly:write. Source and dependency scans additionally require an organization administrator. A DNS scan requires a current ownership proof for the exact authorized domain.

The server returns bounded, organization-scoped findings. It does not expose repository source text, OAuth tokens, user IDs, or private scan cursors. Scan results are potential issues for human review.

MCP tools

Security scans and CVE alerts

The connector has fourteen tools, including the overview card. Read tools inspect project and finding state; write tools start bounded checks or update alert workflow state.

  • Projects and domain verificationlist_projects and get_project show authorized projects and the DNS TXT proof for their domains.
  • Vulnerability reports and scansget_vulnerability_report shows current DNS and source findings. run_dns_scan, start_source_scan, and continue_source_scan run bounded checks.
  • GitHub and dependenciesget_github_connection and get_dependency_scan show connection and scan state. start_dependency_scan checks connected repository lockfiles.
  • CVE alertslist_alerts and get_alert show advisory and fix information. acknowledge_alert and create_alert_ticket update alert workflow state.
  • Overview cardshow_security_overview presents a bounded summary inside supported MCP hosts.

The connector cannot edit a project or its GitHub connection, resolve an alert, or assign a ticket.

CLI

Use ScannerBrightly from a terminal

The scannerbrightly CLI covers projects, vulnerability reports and scans, GitHub connection state, dependency scans, and CVE alerts. Project reads can use an API key; scans and alert operations need a signed-in browser session.

# Inspect the command surface
npx scannerbrightly --help

# Sign in with a browser session to read findings and start scans
npx scannerbrightly login
npx scannerbrightly projects list
npx scannerbrightly vulnerabilities read --help
npx scannerbrightly alerts list --help

The CLI is available as scannerbrightly on npm. Review its source at github.com/scannerbrightly/cli.

Agent Skills

Give your agent the security workflow

The ScannerBrightly skills document the CLI and OAuth MCP security workflows. The security-operations skill explains tool scopes, domain verification, and how to review findings before acting.

npx skills add scannerbrightly/skills

The skills are published at github.com/scannerbrightly/skills. Claude users can also install the ScannerBrightly Claude plugin.

Start building

Build an integration

Register your app, choose its permissions, and submit it for review. Manage registrations for all our products with one developer account.

Register an app · Browse integrations