Bring your agent along.
An agent can help you find relevant work, explore a service, read my résumé, or prepare an inquiry. The tools are available across the public site in browsers that support WebMCP.
Agent tools require JavaScript and a browser with WebMCP support.
Try asking
- “Find Jeff’s mobile application work and the service that fits an app project.”
- “Read Jeff’s résumé and summarize his experience with Laravel.”
- “Find his writing about WebMCP and explain how he approaches an integration.”
- “Help me draft an inquiry about a customer portal. Show me the message before I send it.”
The last prompt prepares an editable draft on this page. You choose whether to open it in your email app and send it. Preparing a draft doesn't send a message or book work.
What the tools can do
fetch_work- Find public projects, client work, and developer tools by keywords or category.
get_services- List services or read one offer’s deliverables, relevant work, and inquiry questions.
get_resume- Read public experience, skills, and selected projects, with a link to the public PDF.
search_writing- Find published articles and their canonical links by topic, or list recent posts.
get_site_info- Get my background, location, contact information, and links around the site.
contact- Get my email address or prepare a visible message draft for you to review.
Work, services, and writing come from the public site collections. The résumé tool uses the same experience and skills as the résumé page. Results link back to their sources.
The pilot planner has three additional tools for preparing a structured project brief. Those are available while you're on the planner page.
Browser support and your data
WebMCP is still evolving. A compatible agent browser must expose the current API for these tools to register. Chrome's WebMCP guide explains its experimental setup; the project's implementation tracker lists browser and agent support.
The site doesn't run a chat model or ask for an API key. Its tools read public information. Contact drafts stay in the current page and are discarded when closed or when you navigate away. If you use an agent, its provider may process the text you share and the tool results it receives.
If your browser doesn't support WebMCP, you can browse every page and email me directly.
Want this in your own application?
I offer WebMCP implementation pilots and broader AI integrations. Start with a useful workflow and test it in the browsers and clients your users need.