Wrap the tools your agent already uses.

Wrap tools passed to `chat.completions.create`. Decide before the function call executes.

Intercept `tool_use` blocks from the Anthropic SDK and Claude Agent SDK before the call executes.

Function-declaration-aware policy evaluation for Google Gemini tool calls.

Authorize Bedrock action groups and Lambda-backed tools before agent execution.

Wrap Cohere tool calls with policy before the model-driven action dispatches.

Evaluate Groq tool calls locally before low-latency agents touch systems of record.

Put policy around Mistral function calls and tool paths before dispatch.

Gate tool calls from open-model workflows running through Together AI.

Streaming-compatible policy checks for Vercel AI SDK tool calls.

Policy at the Mastra tool or workflow step before the governed call runs.

Python guide for wrapping AutoGen tool calls before execution.

Python guide for wrapping CrewAI tools before multi-agent workflows act.

Python guide for checking DSPy modules and ReAct tool calls on the tool path.

Python guide for authorizing Haystack agent tools before data moves.

Wrap LangChain tools and chains with allow, review, or deny decisions.

Put policy on LangGraph ToolNodes and graph context before execution.

Preview typed-tool authorization patterns for PydanticAI agents.

Guide for checking Semantic Kernel plugins before side effects.

Guide for gating browser-agent navigation, form submission, and data access by context.

Policy on Playwright MCP tool calls. Gate navigation, clicks, and downloads by context.

MCP gateway that sits in front of upstream servers and decides each governed tool call against policy.

Native TypeScript SDK. Wrap tool calls at the execution boundary.

Wrap Python callables with local allow, review, or deny decisions.