A chat transcript becomes the interface.
The human has to find the ask, decode the options, and reply in exactly the right format.
LookHere turns AI work into a clear, interactive page. You read, choose, edit, or respond. Then the answer goes back to the exact agent session that needs it.
The person opening the page needs no account, app, or setup.
The agent produces a wall of text. You paste it into another tool, explain what matters, collect an answer, and carry that answer back. LookHere removes that translation work.
The human has to find the ask, decode the options, and reply in exactly the right format.
The context, evidence, open questions, and response controls are arranged around what the human actually needs to do.
Choose a starting grammar below. Then toggle the blocks you want. This is the same composable logic an agent uses when it makes a page.
Put open calls first, keep their context nearby, and let the human answer each one.
No account to create. No new app to learn. No prompt syntax. Try answering this one yourself.
Honest privacy: pages are unlisted but accessible to anyone with the URL. They expire automatically, usually after 30 days.
The final copy is ready. Legal language is unchanged. Mobile QA passed.
The skill teaches your coding agent when a page will help, how to make one, and how to route the response back.
npx skills add alexdobrenko/lookhere-skill --skill lookhere -gThen ask: “Turn this into a page I can respond to.”
Publish structured content to one endpoint and get a shareable URL. LookHere renders the interface, stores draft responses locally, and returns the final response to the workflow that created it.
{
"archetype": "decision",
"blocks": [
"summary", "evidence", "call"
],
"return_to": "originating agent"
}