Sample deliverable

Kavryl AI Agent Permission Risk Report

A realistic example of what security and engineering leaders receive after a Kavryl AI Agent and SaaS Permission Risk Assessment.

Sample company: AcmeCloud Health Exposure score: 74 / 100 Status: High
Executive Summary

AI agents are useful, but their access paths are not yet governed.

Kavryl identified multiple AI agents, MCP servers, and LLM-connected automation paths with access to sensitive SaaS and cloud systems. The highest-risk issue is a support automation agent that can read customer records, access Google Drive documents, and post outbound Slack messages without consistent human approval.

Overall exposure score 74 / 100 High
Key Findings
ID Severity Finding
KAV-001 Critical Support agent can read customer data and post outbound Slack messages.
KAV-002 High MCP server exposes file-system and shell tools in the same agent context.
KAV-003 High Broad OAuth scopes across GitHub and Google Workspace.
KAV-004 Medium RAG app has sensitive document access without strong authorization boundaries.
Agent Inventory
Support Triage Agent Slack, Jira, Google Drive, Salesforce Critical
Dev Workflow Agent GitHub, Jira, AWS High
Internal Docs RAG App Notion, Google Drive Medium
Permission Map
Support Agent
Google Drive
Slack Post
Customer Exposure

The dangerous path is the combination of sensitive read access and outbound communication. A malicious ticket, document, or Slack message could instruct the agent to summarize or forward sensitive customer data into the wrong destination.

Prioritized Remediation Plan

First 7 days

Disable unnecessary shell tools, add approval for outbound messages, remove unused connectors, and assign owners to every agent.

First 30 days

Move agents to least-privilege OAuth scopes, create separate service identities, and add retrieval-time authorization.

First 90 days

Build recurring agent permission review, MCP approval policy, continuous scanning, and audit-ready AI security posture documentation.

Framework Mapping

Findings are mapped to OWASP LLM risks, OWASP Agentic AI guidance, and MITRE ATLAS concepts where relevant. Kavryl uses these frameworks to make agent risk easier to explain to executives, auditors, and engineering teams.

Kavryl Security

Want this report for your AI agents and SaaS permissions?

Book a 20-minute risk review