# Agent access

Public documentation for browser agents, retrieval tools, and future integration boundaries.

## Public surfaces

Browser agents and retrieval tools can read the public product pages, documentation pages, integration status, setup examples, robots.txt, sitemap.xml, integrations.json, llms.txt, and generated Markdown documentation.

The contact form is the public inquiry path for discussing a workflow, demo, or integration requirements.

## Current integration status

Available today: GitHub Issues for work intake and GitHub Copilot CLI for AI execution.

Claude Code (Planned), ServiceNow (Planned), Zendesk (Planned), and other applicable integrations still need to be discussed and approved.

## What agents cannot do here

Authenticated management APIs are not public tools. This public site does not ship a public MCP server, universal connector, public write API, provisioning command, scheduled-demo command, purchase flow, or automatic integration request processor.

A submitted contact request is an inquiry. It is not a purchase, provisioning action, scheduled demo, or successful integration.

## Form interaction expectations

Contact form submission must be an explicit user action. Page loads, crawler visits, or prefilled URL parameters must not submit the form.

The form uses visible labels, stable field names, and explicit success or error states for humans, assistive tools, and browser agents.

## Future MCP or OpenAPI gate

A future public MCP or OpenAPI project needs a real customer workflow, confirmed read and write operations, scoped authentication, tenant boundaries, abuse controls, and an approved implementation plan before any connector is published.

## Discovery file expectations

llms.txt and the AI-oriented Markdown or JSON files are optional navigation aids for agents and non-Google consumers. They are not Google Search ranking factors, indexing requirements, or guarantees that any AI provider will use the content.