Understand what your AI tools can really do.

Analyze MCP servers, OpenAPI specifications, Claude tools, Codex tools, and other callable AI capabilities to identify permissions, trust boundaries, excessive agency, and attack surface before attackers do. Tool Capsule Analysis performs static analysis of tool definitions; it does not execute tools.

Analyze Tool Definitions View Sample Report

WHAT CAN YOUR AI TOOLS REALLY DO?

Discover

Inventory every callable capability from MCP, OpenAPI, Swagger, Claude, Codex, Gemini, and agent frameworks.

Understand

Classify permissions, authentication, side effects, data flows, and trust boundaries.

Prioritize

Identify excessive agency, missing approvals, destructive operations, and risky cross-boundary tools.

Feed evidence

Generate Threat Canvas assets and feed Agent Permission Diff and Correlation Engine workflows.

Core capabilities

What SecEng Tool Capsule Analyzer does.

Tool Inventory

Build a callable-tool inventory across MCP servers, OpenAPI specifications, Swagger files, Claude tools, Codex tools, Gemini tools, LangGraph, CrewAI, AutoGen, Semantic Kernel, Flowise, Langflow, and n8n.

Capability Classification

Classify read, write, delete, send, execute, admin, filesystem, browser, secret, and network capabilities from tool definitions and API schemas.

Permission Model

Map permissions, authentication requirements, inherited capabilities, and side effects so tool scope is reviewable before deployment.

Trust Boundary Mapping

Identify tools that cross trust boundaries, expose destructive operations, or lack approval gates for irreversible actions.

Downstream Handoff

Produce structured findings that feed Threat Canvas, Agent Permission Diff, and release-readiness review.

Static Analysis First

Analyze definitions and schemas without invoking production tools, calling APIs, or changing customer systems.

Evidence & signals

What you get out of the box.

Inputs

MCP servers
OpenAPI / Swagger
Claude and Codex tools
Gemini tools
LangGraph / CrewAI / AutoGen
Semantic Kernel / Flowise / Langflow / n8n

Security Model

Tool inventory
Capability graph
Permission model
Authentication requirements
Side effects
Trust boundaries

Deliverables

Threat Canvas assets
Engineering findings
Executive summary
Permission drift handoff
Release review signals

AI SECURITY ENGINEERING WORKBENCH

Ready to put SecEng Tool Capsule Analyzer to work?

Tool Capsule Analysis is an active-development SecEng Workbench capability available through scoped public-site review conversations. We analyze tool definitions, model permissions, and return engineering findings without executing production tools.

Start AI Security Assessment See the full Workbench

View pricing →

Also in the Workbench

WHAT AI DO WE HAVE?

SecEng Surface Scanner

Browser, Repo & IDE AI Discovery