タグ：ai-security

Screens an AI agent's tool calls before they run and tool results before the agent reads them. An MCP proxy plus a hooks adapter for a client's built-in tools.

Open-source MCP and HTTPS screening for prompts, skills, and tool descriptions using TypeSafe AI Jev, deterministic checks, and configurable reasoning.

Recorded comparison of Jev, Terra, and Opus on 100 synthetic security-triage cases, five passes each, with a static inspectable dashboard

Static code scanner (SAST linter) that finds insecure usage of Jev / TypeSafe guardrails in Python. Not a guardrail itself — build-time analysis, offline, no API key.

Typed answers with a probability for agent stacks: a choice, a score or a yes/no from a versioned template. Only the fields a template names leave the box, and every answer is on the record. An optional add-on to the TAIPANBOX stack, over HTTP and MCP.