When your coding agent needs your judgement, it hands you one offline page instead of a wall of chat. Answer, skip, add what it missed, send one file back. It checks your answers before it acts, and round 2 shows what changed. Free, open, no account.
Hi Product Hunt, I'm the maker.
I built this because every coding agent I use asks for my judgement the same way: a long chat message. I answer the parts I notice, "looks good" covers the rest, and nothing is written down.
LetMeShowYouSomething is a skill for your agent (Claude Code, Codex, Hermes). When it needs you to judge several things, it writes them down as a review, checks it, and hands you one page:
• One question at a time, with the screens it proposes and a map of the flow beside each
• Skip what you can't judge yet; add what it forgot to ask
• Send back one file. The agent checks it adds up and starts its report with what is still open. An unanswered question is never taken as a yes
• Round 2 shows what changed after your notes, before and after
The page is one HTML file that loads nothing, so the person who decides needs no account and no internet.
It was built with Claude Code, and it is free and open source. The tutorial shows one real run, a salon's booking app, from idea to round 2.
I'd love to hear what you would want an agent to ask you this way.