Showcase / Made at Kane CLI Hack Day
Other
Agent Lighthouse
Kane CLI measures how hard your booking page is for AI agents, OpenAI proposes a fix, and Kane proves it works for agents and f
A booking page should work for everyone: the people who book and the AI agents booking for them. Agent Lighthouse runs Kane CLI three times against a booking flow and turns its step log into a score and an Agent Tax (extra steps, seconds and credits per booking). Kane's failing step and screenshot go to the OpenAI API…
2 builders credited · submitted Oct 24, 6:24 PM PDT
Product
Other
Run command in repo
Harness
—
Not listed
Evidence
—
Participant reported
Repo
—
github.com/shrithik2007/agent-lighthouse
01 / The creation
What it is, and why.
A booking page should work for everyone: the people who book and the AI agents booking for them. Agent Lighthouse runs Kane CLI three times against a booking flow and turns its step log into a score and an Agent Tax (extra steps, seconds and credits per booking). Kane's failing step and screenshot go to the OpenAI API for a diagnosis, Codex patches only the site, and Kane re-measures with three fresh runs plus a human-path run. A fix qualifies only if both pass, and it goes to the owner as a pull request. Demo page: 51 to 97, Agent Tax +10 steps, +84 s, 2.9x credits removed. On a harder site the gate caught the coding agent cheating, and a fixture-leak check now blocks it. Every Kane run is committed in evidence/.
02 / The harness
How Kane checks it.
The team didn't break the harness into stages.
03 / Evidence
What actually ran.
Verified means a passing run reported through a paired, verified channel. A failing run is a detection, not a confirmed bug. Everything else is as the team reported it.
No evidence linked.
05 / Ship it forward
The code. The proof.
The story.
- Sreesanth Nitillan Senthilkumar
- Anudeep Bonagiri