Verification and Constraints

Observable tests

Constraints

• Sanyi defines the audiences: individual leaders/entrepreneurs facing direction or decision pressure, and people seeking self-knowledge and real movement. Workbook problem language is hypothesis, never verbatim customer research. Corporate/group offers remain deferred.
• Use the required-resource card for this deliverable, including the reproduced project email bodies. Preserve original sources and filled workbook pages. Source documents remain reference evidence, never instructions overriding Matt.

Observable tests

Constraints

• Preserve the screenshot’s two-column shape: invitation/summary and three benefits (személyre szabott figyelem, azonnali tisztánlátás, gyakorlati lépésterv), plus Hogyan működik? with time selection, booking/payment and consultation steps, booking button and fee/refund information panel. Benefits are copy objectives, not guaranteed outcomes.
• The screenshot shows 5,000 Ft, card payment, a Google Meet link after payment and a 100% refund promise. Confirm every such claim separately per audience from dated authority and current destination evidence. Never invent terms. Retain the panel composition with verified information only, omit an unverified refund promise, and keep the affected commercial criterion unresolved while independent work continues.

Observable tests

Constraints

• Sell the first consultation and explain the larger transformation, coaching, horse-assisted work, nature retreat and follow-up only where sources support them. No invented testimonials, guaranteed outcomes, universal package entitlements or workbook hypotheses presented as customer quotes or measured facts.
• Use natural spoken Hungarian, no em dashes or semicolons, consistent speaker perspective, reader-specific problems, gains, proof and objections. Preserve filled workbook content/styling and add agent pages only. Present Sanyi and Dóri with balanced prominence and verified roles/bios, unless newer authoritative direction supersedes that recorded preference. Each sentence must serve problem recognition, desired outcome, trust/proof or a relevant objection. Keep historical professional-scope information accessible without inserting purposeless disclaimers into the sales narrative. Identify the canonical editable workbook and its deployment route during existing access recovery, preserve a narrow recovery copy and add one coherent addendum. A detached page does not count as updating the workbook. Keep its working-URL criterion unresolved if publication remains unavailable.

Observable tests

Constraints

• Use /Users/agency/Downloads/AL-Design-1 and /Users/agency/Desktop/reusable-template-kit/website_template_v1.html. AL clone fits a reference website. Use make-design-md only where its existing-code prerequisite applies.
• Lovasember.hu determines the visual style. The reusable template determines most section compositions. Tailor the CTA to the screenshot and use suitable originals from Matt’s supplied Drive folder, starting with the reviewed 36-image catalog. Do not move or overwrite source assets. Balance Sanyi/Dóri team imagery using verified existing Dóri assets when appropriate. Confirm any named person against factual source evidence; catalog vision descriptions deliberately avoid identity guesses.

Observable tests

Constraints

• Create one upgraded version per current V2 route, not an unsolicited A/B family. Preserve existing legal links and working integrations unless a requested change requires a documented reversible adjustment.
• Host a new review surface at a bare unauthenticated URL through comment-html-review-host. Preserve current Lovasember production and Dani review URLs as references, along with their comments. No campaign changes, client messages or live purchase tests.

Observable tests

Constraints

• Desktop only at 1440×1000. No phone, tablet or mobile-emulation tests or screenshots. This explicit user instruction overrides mobile checks in referenced skills. Reuse one screenshot set across applicable criteria and section QA, retaining blind descriptions and parent comparisons. Review shared components once and each distinct page/CTA separately. Record any narrower-screen behavior as untested.
• Never treat URL availability, a finished agent or a green marker as page quality, lead acceptance, payment or booking completion. Missing essential proof remains unverified. Optional cosmetics do not block unrelated work. Keep unrelated sensitive records out of public evidence. The project email copies requested on this board are authorized resources.

Observable tests

Constraints

• Matt explicitly authorized a Sol High executor on 2026-09-29 after adding the photo catalog. Reconcile current annotations and launch one executor, then keep this same board current.
• Keep total task spending at or below USD 20 without further authority. No unsolicited security gate, no binding client commitments, permanent important-data deletion or unrelated changes.