# SecondDoor: 11 point Door Gap
- Site: seconddoor.io
- Scanned: 2026-09-01
- Scan type: Basic scan
- Business type: hybrid
- Report: https://www.seconddoor.io/r/seconddoor-s3v8h5
- Scanner version: 8
A storefront and SaaS platform with strong machine discoverability but no published path for agents to complete purchases.
## Scores
| Measure | Score |
| --- | --- |
| Humans | 83/100 |
| Agents | 72/100 |
| Composite | 78/100 |
| Door Gap | 11 points |
The human score is the mean of the four human dimensions, the agent score the mean of the six agent dimensions, and the Door Gap the difference between them.
## What the machine actually did
The agent got through 13 of 14 requests, but could not find out whether verified agents are recognised.
- ok      reach the site at all
- ok      find out whether machines are welcome
- ok      ask whether the door opens for requests carrying the AI crawlers' names
- ok      read the site's own guide for language models
- ok      find out what pages exist
- ok      find a contact route for machines
- ok      find a published agent interface
- ok      find an agent plugin manifest
- missing find out whether verified agents are recognised
- ok      read the homepage as data rather than as a page
- ok      find a machine readable API spec
- ok      read the developer documentation
- ok      read the prices without running JavaScript
- ok      walk the way a visitor would, to a signup form or a product
## The agent door
- **machine_discoverability** 100/100. Robots.txt names 14 AI crawlers and allows them all. A real browser is served the page normally. No user agent rule refuses declared crawler names. Sitemap lists 59 URLs. Homepage carries 6 schema.org types. llms.txt is published.
- **programmatic_onboarding** 50/100. A machine readable API spec is published.
- **pricing_legibility** 85/100. 6 prices are readable without JavaScript on the pricing page: $0, $29, $99, $249, $2,500, and one more. No tier is gated behind a sales contact. Prices are legible to machines and humans alike.
- **agent_aware_instrumentation** 100/100. MCP manifest published at /.well-known/mcp.json. AI plugin manifest published at /.well-known/ai-plugin.json. Robots.txt names 14 AI crawlers and allows them. OpenAPI spec published.
- **machine_fetchable_trust** 85/100. Security.txt is published at /.well-known/security.txt with a contact route for machines. The homepage answers automated requests with a 200 status in 474ms. Trust signals are present and accessible.
- **commercial_rails** 10/100. No UCP manifest at /.well-known/ucp. No commerce platform or payment processor detected on the homepage. No published route for agents to complete a purchase. An agent can discover the product and read the prices but cannot transact.
## The human door
- **steps_to_first_value** 80/100. The primary call to action, 'Sign up for free', is 2 steps from the homepage and leads to a 2-field signup form. A human can reach the form quickly. What happens after submission was not observed.
- **required_fields** 93/100. The signup form has 2 fields, both required: Work email and Password. Google single sign-on is offered as an alternative. 2 of 2 inputs carry validation rules. The form is minimal and direct.
- **verification_walls** 80/100. No verification requirement appears on the signup page itself. Google single sign-on avoids email verification entirely. The form does not block progress with a second factor or email confirmation step.
- **error_recovery** 80/100. Both inputs on the signup form carry validation rules, so a user can correct mistakes before submission. Inline error handling was not observed on the page, but validation attributes are present.
---
Scored by SecondDoor. Method: https://www.seconddoor.io/methodology
Scan another site: https://www.seconddoor.io