PDF · 9 PAGESCore white paper
Securing AI Agent Action Paths
A security model for governing agents across MCP, APIs, SaaS connectors, browsers, code execution, workflows, and delegated agents. Covers visibility, policy-based prevention at supported control points, containment, remediation, verification, JADEPUFFER boundaries, and a staged product path.
MCP & other routesPrevention boundariesContainment & remediationNIST AI RMF
PDF · 6 PAGESTechnical architecture
AI Agent Security Architecture
A reference architecture separating Pantheon’s proposed control plane from the enforcement points and existing customer controls that carry out decisions. Includes an action-path coverage model, event data, response sequence, integration examples, validation gates, and security requirements.
Reference architectureAction-path coverageIntegrationsPrototype validation
PDF · 4 PAGESBuyer overview
Evaluating Pantheon Control
A concise guide for security leaders and technical evaluators: the problem the concept addresses, proposed workflow, route coverage, fit with existing controls, pilot steps, and questions to ask before a purchase decision.
Buyer fitUse casesPilot criteriaClaim boundaries
PDF · 5 PAGESTechnical guide
Integration, Data & Deployment Guide
A practical discovery and validation guide covering MCP, APIs, SaaS, browser, shell, CI, workflow, and delegated-agent paths; proposed event fields; data-handling requirements; permission tiers; failure modes; and a staged pilot.
Integration catalogData handlingResponse governanceValidation plan