Developer tools

Browser-side tools for agentic integration

Simulate agent governance policies, lint MCP tool definitions, draft A2A Agent Cards, and format or transform integration payloads. Tool inputs are processed locally in your browser.

Privacy: Tool inputs are processed locally in your browser; the tools do not upload what you paste or type. Avoid production credentials or confidential enterprise payloads.

New · interactive

Agent Policy Simulator

Test how governance policies decide what an AI agent may do: pick a scenario such as a large payment, a bank-detail fraud attempt or a prompt injection, and see Allow, Deny or Approval required with the rule, trace and audit record. Edit the policy and watch the decision change.

Open the simulator →

MCP tool definition linter

Paste one tool object (or an array, or a tools/list result). Checks naming, descriptions, JSON Schema hygiene and safety annotations against enterprise practice and the MCP 2026-07-28 specification. Advice, not certification.

Tool definition JSON

Findings

Findings appear here.

A2A v1.0 Agent Card builder

Generates a draft card for validation.

Fill in the basics to generate a starting Agent Card in A2A v1.0 shape. Add security schemes and signatures for your environment, and validate against the official A2A schema before publishing.

Agent name Version Description Organization Organization URL Endpoint URL Protocol binding Security Streaming Push notifications Extended card for authenticated callers Skills (one per line: id | name | description | tags,comma,separated)

Agent Card JSON appears here.

Payload utilities

Paste sample JSON, XML or text

Result

Output appears here.

XPath mode input: XML document, then a line containing ---XPATH---, followed by an XPath expression. XML checks use the browser XML parser; test against your target runtime.