Security and commercial governance should be specific enough to inspect.

This page describes the safeguards that prevent generated interpretation, draft capability or unapproved constraints from becoming an unsupported commercial commitment. For infrastructure and data-handling detail, ask us directly and we will answer in a security review.

Commercial safeguards

What stops a plausible answer becoming an approved offer.

  1. Draft Capability Maps require owner validation

    Generated or imported capability is not automatically treated as a statement of what the business can deliver.

  2. Constraint Records require approval

    Draft constraints must identify their evidence basis, owner and approval status before they can govern an offer.

  3. Prices require an approved basis

    Specify does not invent price logic or present generated estimates as approved prices. The analyzer is explicitly forbidden from producing a price, cost, lead time or delivery date.

  4. Delivery commitments require evidence

    A requested date remains a request until the applicable Constraint Record is satisfied and the responsible owner confirms capacity.

  5. Technical conclusions require specified evidence

    Technical approval must reference the drawings, dimensions, materials or other evidence defined by the business.

  6. Decision Records remain attributable

    Approved, revised, referred and rejected outcomes remain connected to the applicable constraints, human reviewer and reasoning.

  7. Opportunity Ledger access is controlled

    Unsuccessful and incomplete demand may contain commercially sensitive information and is visible only to authorized users.

Responsible disclosure

To report a security concern, contact the founders directly. Include a description of the issue and the steps required to reproduce it. Do not access or modify data that does not belong to you.

A dedicated, monitored security address will be published here before launch.

Engineering detail

How the system is architected

The technical page sets out the design principles, the production architecture and the boundary between AI interpretation and commercial authority.