Continuous Monitoring

Sentry: Production AI agents, MCP servers, and applications with privileged tools or customer data.

Monitor the tools and permissions behind your AI. Test authenticated agents and MCP servers for changes that expose data or expand access.

Included

Monitoring coverage

  • Everything in Patrol, with unlimited bot, agent, or MCP endpoints
  • MCP tool inventory, schema changes, and permission drift monitoring
  • Daily adaptive prompt injection, tool and agent poisoning tests
  • Authenticated journeys, role boundaries, and token revocation probes
  • 25 canary phrases and 100 business facts across the plan
  • RAG leakage, memory poisoning, and approved tool misuse scenarios
  • Webhook, Slack, and email alerts with traces and fix tasks
  • Managed CI/CD regression corpus and quarterly security evidence pack
Plan limits

Included endpoints and cadence

  • Unlimited endpoints, including MCP servers
  • 25 canary phrases and 100 business facts per plan
  • Hourly health checks and daily adaptive adversarial tests
  • Authenticated and MCP checks use customer-provided test access and approved tools
Outputs

What your team receives

  • MCP inventory and tool, schema, and permission change alerts
  • Attack traces with the affected role, tool, and access boundary
  • Regression tests and Markdown remediation tasks
  • Monthly reports and a quarterly security evidence pack
Setup

From intake to first signal.

Connect your endpoints, define protected facts and canaries, and approve the testing schedule. Sentry setup also confirms MCP tools, permissions, and test credentials.

Connect your bot or MCP server

Share your bot, API, or MCP endpoint. We confirm compatibility, test accounts, tool permissions, and the checks included in your plan.

Add facts and canaries

Define the business truth set and place canary phrases where the bot can see them but should never reveal them.

Run external probes

WardenBot drives the bot from the outside on the tier schedule: heartbeat, probe packs, multi-turn tests, and journeys.

Plan fit

Sentry is the right fit when...

Production AI agents, MCP servers, and applications with privileged tools or customer data.

Other monitoring tiers

  • Watch $299/mo - For everyday monitoring
  • Patrol $499/mo - For growing AI products
Ready for proof

Start Sentry monitoring intake.

Share your endpoint, protected facts, test access, and alert preferences. We confirm compatibility and scope before recurring tests begin.