Directory
‹ Back to apps

Vulnify

Runtime authorization for AI agents. Before Cursor's agent reads, writes, deletes, exports or sends data, ask Vulnify and get ALLOW, REVIEW or BLOCK with a risk score and reasons. Sign in with Vulnify (OAuth), no API key to paste.

What’s inside

1 MCP server · 2 skills
  • Skillvulnify-guard-actionsWhen to call check_action, decide_action or get_decision before a sensitive action, and how to handle REVIEW.
  • Skillvulnify-policies-as-codeList, test and plan Vulnify policies as code without writing them.
  • MCP servervulnifyHosted Vulnify MCP server. Sign in with Vulnify (OAuth).