WisPear: community wisdom, whispered to your agent

WisPear came out of ETHGlobal Cannes, built with a team of five. The one-line pitch is the tagline we landed on: community wisdom, whispered to your agent.

The problem we picked

Everyone is suddenly building agents, and the hard part isn't the model. It's the tools. Which tool actually does what you need? Which one is good today, not six months ago? Right now that knowledge lives in scattered threads and gut feeling.

WisPear is a marketplace and discovery layer for agent tools, built on Intuition. You describe what you want your agent to do ("find me job offers") and it returns the tools the community recommends for that exact context, ranked by what performs best, plus a ready-to-paste meta-prompt so your own AI builds the thing the way you want it. Because tool quality is recorded as Intuition attestations, the recommendations are community-verified instead of guessed. It's the atom/triple idea from Sofia, pointed at tools instead of people.

Building as a team of five

Solo, I can hold the whole thing in my head and move fast. With five people under a hackathon clock, the bottleneck moves from typing to agreeing, fast. The win was carving the work into clean seams (the attestation/data side, the end-user flow, the agent/meta-prompt piece) so five of us weren't editing the same files at 3am. Keeping everyone pointed at the same demo was as much of the job as the code.

What stuck with me

The meta-prompt output is the part I keep thinking about: the deliverable isn't "a tool," it's the instructions to build the right tool, verified by people who already did it. That feels like the actual shape of how agents and humans will share knowledge. WisPear was a weekend's version of it, but the direction is the keeper.