body { margin: 0; font-family: var(--cf-font-body, "Manrope", sans-serif); color: var(--cf-text, #1b1b1b); background: var(--cf-bg, #ffffff); /* Design system §2: body line-height 1.5, --text-body 16px. */ line-height: var(--cf-line-height, 1.5); font-size: var(--cf-font-size-base, 16px); } /* Padding counts INSIDE a declared width, everywhere. Without this, any element carrying both a width and padding resolves wider than it asks for. Measured on a real template whose root had .cf-section and .cf-shell on the SAME element: 980px of content plus 24px of gutter each side computed to 1028px in a 390px viewport, and the section overflowed. Scoped to the generated site only, so the admin UI's own styling is untouched. */ .cf-site *, .cf-site *::before, .cf-site *::after { box-sizing: border-box; } /* Nothing may be wider than the VIEWPORT. ------------------------------------------------------------------ The editor bakes an element's COMPUTED desktop size into an id rule when it is resized or restyled, e.g. an id rule setting width:900px. An id selector outranks every class here, so a card's own text was 900px inside a 358px card and simply got clipped. Measured on a real page, twice: 38 such rules, then 39 more created by later edits. A repair script fixes what exists; this stops the next edit from breaking the phone again. Still true across the library: 204 such id rules, 148 of them wider than a 390px phone and 99 at 900px or more — so this guard cannot simply be dropped. WHY THIS IS BREAKPOINT-SCOPED RATHER THAN UNCONDITIONAL. :where() contributes no specificity, so this rule should lose to any real declaration — including a Tailwind utility such as max-w-[1280px]. It did not: STARTER_STYLE is injected once per custom-component block and therefore lands AFTER the framework stylesheet, and at equal (zero) specificity the later rule wins on source order. Every deliberate max-width on such a page was being flattened to 100%. Measured on a live site: a collection section whose container asks for 1280px rendered at the full 1440px viewport width, edge to edge, while the same block on a page with no custom-component blocks — and therefore no copy of this stylesheet — rendered correctly. Confirmed by disabling the injected copies in the browser: the container returned to 1280px immediately. 8 of 12 tenants and 100 of 145 pages carry such blocks, so this was platform-wide rather than one tenant's problem. Limiting the clamp to viewports narrower than the design system's container keeps exactly the protection it was written for — a 900px element on a phone — while leaving desktop layout to the authored CSS, which is where a 1280px container is meaningful and where no clamping was ever needed. THE EDITOR'S DESKTOP FRAME IS EXCLUDED. The canvas is only ~1115px wide - narrower than this 1280px breakpoint - so the editor sat permanently inside the clamp and capped every element at 100% of its parent, even on the DESKTOP device. Setting a width above the canvas width therefore appeared to do nothing: the Style Manager wrote it correctly (verified: model width 459px -> 333px) and the canvas silently overrode it. Measured: an element asking for 1600px renders 1600px on a 1900px live page and 1067px in the canvas. The exclusion is scoped to the DESKTOP frame only, by raising the breakpoint for the canvas to the tablet width. The editor's 375px and 768px frames are genuinely narrow and must still clamp - a phone preview that does not show the overflow a real phone will show is worse than useless. Excluding the canvas outright was tried and reverted for exactly that reason: it let a 1600px element render 1600px inside a 375px frame. */ @media (max-width: 1280px) { body:not(.cf-editor-canvas) :where(*) { max-width: 100%; } } /* The canvas keeps the guard at real device widths. 992px is the tablet media width, so both narrow frames are covered while the desktop frame (~1115px) is not. */ @media (max-width: 992px) { body.cf-editor-canvas :where(*) { max-width: 100%; } } /* Below the tablet breakpoint a page must never scroll sideways, and this is the one clamp allowed to beat the page's own rules. WHY !important IS NEEDED. Both clamps above use :where(), so they carry zero specificity and are a pure fallback - deliberate, because a stronger desktop clamp was tried and it overrode all five header templates' own nav spacing. But a section rule like .component-x .contact-d2v1-head is (0,2,0) and beats a fallback every time. Matching that specificity does not help either: the page's CSS is injected AFTER this sheet, so a tie goes to the page. Measured on maiitrii at 390px - clamp variants that were tried and did NOT reflow the element: .cf-site :where(...) (0,1,0) 585px - lost on specificity .cf-site .cf-site :where(...) (0,2,0) 585px - lost on source order overflow-x: hidden on the wrapper was also tried: it stops the sideways scroll but leaves the element 585px, so the content is merely clipped off-screen and unreachable. Only !important actually reflows it (358px). THE CAUSE THIS CATCHES. max-width in ch resolves against font size, so a section with a large font turned max-width: 58ch into 585px inside a 390px phone - the page rendered 601px wide and everything looked shifted. There are ~530 ch widths across the templates, and whether any given one overflows depends on the tenant's type scale, so this is a floor rather than a fix to any single value. DESKTOP IS UNTOUCHED - the breakpoint stops at the tablet width, so the nav regression above cannot recur. table/pre are left out so the overflow-x rule below still governs them, and .cf-allow-wide opts a genuinely full-bleed element back out. */ @media (max-width: 992px) { .cf-site :where(section, div, p, span, h1, h2, h3, ul, figure, img):not(.cf-allow-wide) { max-width: 100% !important; } /* An explicit WIDTH beats max-width, so the rule above cannot rescue a box that was resized in the editor. --------------------------------------------------------------------- The editor bakes a resize into an #id rule as a fixed pixel width. Measured on maiitrii/services at 390px: #igikyj{width:530px} inside a correctly-clamped 342px parent - max-width:100% resolved to "100% of my parent" and changed nothing, so the box overflowed by 188px and the page scrolled sideways at 554px. 39 such rules exist across the stored pages, including 1280px and 1284px ones. Releasing the width lets the box take the space it actually has. Applied only below the tablet breakpoint, where a fixed pixel width is never what anyone wants. ELEMENTS THAT ARE MEANT TO HAVE A SIZE ARE EXEMPT. A logo, avatar, icon or badge carries an explicit width deliberately; flattening those was measured and rejected - width:auto on everything collapsed a 150px logo to 40px and stretched a 40px avatar to the full 342px. is excluded for the same reason, and .cf-keep-size is the escape hatch for anything else that must hold its size. THE LAYOUT SHELL IS EXEMPT TOO. .cf-shell / .cf-container are the page's width wrappers: they carry width: 100% with margin-left/right: auto so content spans its column and centres. width:auto on a FLEX ITEM makes it shrink-to-fit, and the auto margins then split the leftover space - so the wrapper collapsed to its content and sat centred with dead space either side, instead of spanning the bar. Measured on maiitrii's header at 390px: .cf-shell computed to 230px with 64px margins, leaving the logo 80px from the left edge and the hamburger 80px from the right, bunched toward the middle. Opening the nav moved them to 24px and 20px, so the two states did not line up. Reported as the logo and hamburger not stretching to match the open nav's width. Verified that exempting them does not reopen the overflow this rule exists for: with the exemption a 530px fixed-width child still clamps to 390px and the page does not scroll sideways, while the shell goes from 212px to 358px and the edge gaps from 89px to 16px. SCOPED TO ELEMENTS THAT CAN ACTUALLY CARRY ONE. The baked width always arrives as an id rule - #igikyj { width: 530px } - so only an element WITH an id can ever have one. Without [id] this rule fired on everything. Measured on aarya, a tenant whose pages are hand-written Bootstrap rather than editor-built: the unscoped rule matched 2163 elements on /financial-services and changed the computed layout of 2288 of them, 290 on the home page and 205 on /about-agc. That site has ZERO #id width rules on any page, so every one of those changes was collateral — it was reshaping a Bootstrap grid to defeat a problem that did not exist there. Reported as cf-site disturbing the width on every page. Releasing the width is still not what fixes overflow: removing this rule entirely changed the document width on none of five pages, and /about-agc overflows at 405px in a 375px viewport both with and without it. The max-width clamp above is what holds the page. [id] keeps the rule pointed at the real case (an editor-resized box) and off markup it was never written for. */ .cf-site :where(section, div, p, h1, h2, h3, ul, figure)[id]:not(.cf-allow-wide):not(.cf-keep-size):not(.cf-shell):not(.cf-container):not([class*="avatar"]):not([class*="logo"]):not([class*="icon"]):not([class*="badge"]) { width: auto !important; } } /* Table and pre need to scroll rather than shrink, or their content becomes unreadable instead of merely narrow. */ .cf-site :where(table, pre) { display: block; overflow-x: auto; } .cf-section { /* §5: section padding --space-80 desktop, stepping down at the tablet and phone breakpoints below. The horizontal gutter follows §4 (24px). */ padding: var(--cf-section-padding-top, 80px) 24px var(--cf-section-padding-bottom, 80px); } /* Scroll jank on large pages, worst on hand-coded ones that never adopt .cf-section/.cf-shell at all (a Tailwind/Bootstrap import, not a page built from the block library — .cf-site is the only wrapper such a page actually has). The max-width clamp two rules up matches every section, div, p, span, h1-h3, ul, figure and img on the page — measured on one real tenant page at 2163 elements. Each sits under a :where() (zero specificity) plus one !important, so on any reflow the browser still has to walk the full subtree to resolve them, and a resize/orientation change — which the clamp exists to react to — forces that walk again on a page this dense. CSS contain (layout + style) isolates a subtree's internal layout and scoped counters/quotes from the rest of the page, so a change inside it no longer forces the header, footer or other top-level blocks to be re-checked too. Scoped to

— the wrapper every page route (src/app/[domain]/page.tsx and .../[slug]/page.tsx, every branch) already renders around the page's own content, and nothing else — because that is real page content and nothing chrome ever needs to escape: - [data-cf-header] uses position:fixed for tenants with headerOverlapsContent (data-cf-header-overlap="true" on .cf-site itself). Containment creates a new containing block for a fixed/absolute descendant, so containing the header would pin it to its own box instead of the viewport the moment the page scrolled. - SiteFooter renders no wrapper of its own (a bare fragment), so there is no single element to safely contain without also risking whatever position:fixed content a footer block might hold. - The mobile tab bar (MobileTabBarBlock) is portaled straight to document.body and is never inside .cf-site at all.

avoids all three: it is real, present on every route, holds only the page's own sections, and was never given fixed positioning. Unlike content-visibility: auto, contain does NOT skip rendering or collapse the box for off-screen content, so it cannot break the on-scroll IntersectionObserver animations (grapesAnimations.ts) or any getBoundingClientRect() measurement of a section's children — those keep seeing real, current layout. */ .cf-site > main { contain: layout style; } .cf-shell { /* §4: container max-width 1280px; larger viewports gain outer margin rather than wider content. */ max-width: var(--cf-container-max-width, 1280px); margin: 0 auto; /* max-width alone does not stop a wide CHILD pushing the container past the viewport — a grid whose tracks resolve larger than the screen widens this box, and the page then scrolls sideways. Measured on a phone: the shell reached 980px inside a 390px viewport. width + min-width:0 pin it to the available space and let its grid children shrink instead. */ width: 100%; min-width: 0; } .cf-kicker { display: inline-flex; align-items: center; gap: 10px; font-size: 12px; font-weight: 700; letter-spacing: var(--cf-letter-spacing, 0.12em); text-transform: uppercase; color: var(--cf-primary, #0b6b57); } .cf-title { /* §2: H1 48px via clamp(32px, 4vw, 48px) so it scales fluidly instead of jumping at breakpoints; heading line-height 1.25 and -0.02em tracking. */ margin: 18px 0 14px; font-size: var(--cf-font-size-h1, clamp(32px, 4vw, 48px)); line-height: 1.25; letter-spacing: -0.02em; font-family: var(--cf-font-heading, "Instrument Serif", serif); font-weight: var(--cf-font-weight-heading, 400); text-transform: var(--cf-text-transform-heading, none); color: var(--cf-heading, inherit); /* A long brand word must not push the whole PAGE wider than the phone. Measured on a real site: "ChangeMantras" in a heading was 388px inside a 358px column, which widened the document to 405px on a 390px screen — the page scrolled sideways and a strip of it showed beside the open mobile menu, which read as the nav being cut off. Only applies when a word genuinely cannot fit, so normal headings are unaffected. */ overflow-wrap: break-word; } .cf-body { font-size: var(--cf-font-size-base, 16px); line-height: var(--cf-line-height, 1.5); color: var(--cf-text, #4a4a4a); } .cf-button { display: inline-flex; align-items: center; justify-content: center; padding: var(--cf-btn-py, 14px) var(--cf-btn-px, 22px); /* §6: buttons/inputs sm–md. */ border-radius: var(--cf-btn-radius, 8px); background: var(--cf-btn-bg, #0b6b57); color: var(--cf-btn-text, #ffffff); font-weight: 700; text-decoration: none; font-size: var(--cf-btn-font-size, 16px); border: 1px solid var(--cf-btn-border, transparent); } .cf-button:hover { background: var(--cf-btn-hover-bg, var(--cf-btn-bg, #0b6b57)); color: var(--cf-btn-hover-text, var(--cf-btn-text, #ffffff)); border-color: var(--cf-btn-hover-border, var(--cf-btn-border, transparent)); } /* Secondary Button: Outlined style using theme secondary/primary color */ .cf-button.cf-button-secondary, .cf-button-secondary, button.cf-button-secondary, a.cf-button-secondary { background: transparent !important; background-color: transparent !important; color: var(--cf-btn-secondary-text, var(--cf-secondary, var(--cf-btn-bg, var(--cf-text, #0b6b57)))) !important; border: 2px solid var(--cf-btn-secondary-border, var(--cf-secondary, var(--cf-btn-bg, var(--cf-border, #0b6b57)))) !important; box-shadow: none !important; } .cf-button.cf-button-secondary:hover, .cf-button-secondary:hover, button.cf-button-secondary:hover, a.cf-button-secondary:hover { background: var(--cf-btn-secondary-hover-bg, rgba(0, 0, 0, 0.06)) !important; background-color: var(--cf-btn-secondary-hover-bg, rgba(0, 0, 0, 0.06)) !important; color: var(--cf-btn-secondary-text, var(--cf-secondary, var(--cf-btn-bg, var(--cf-text, #0b6b57)))) !important; border-color: var(--cf-btn-secondary-hover-border, var(--cf-secondary, var(--cf-btn-bg, var(--cf-border, #0b6b57)))) !important; } /* Link Button: Clean text link style */ .cf-button.cf-button-link, .cf-button-link, button.cf-button-link, a.cf-button-link { background: transparent !important; background-color: transparent !important; border: none !important; border-color: transparent !important; padding: 0 !important; min-height: auto !important; color: var(--cf-link, var(--cf-primary, var(--cf-btn-bg, #0b6b57))) !important; text-decoration: underline !important; box-shadow: none !important; } .cf-button.cf-button-link:hover, .cf-button-link:hover, button.cf-button-link:hover, a.cf-button-link:hover { background: transparent !important; background-color: transparent !important; color: var(--cf-link-hover, var(--cf-link, var(--cf-primary, var(--cf-btn-bg, #0b6b57)))) !important; text-decoration: underline !important; } .cf-grid { display: grid; gap: var(--cf-row-gap, 24px) var(--cf-column-gap, 24px); } .cf-grid-2 { grid-template-columns: repeat(2, minmax(0, 1fr)); } .cf-grid-3 { grid-template-columns: repeat(3, minmax(0, 1fr)); } .cf-card { padding: 28px; /* §6: cards --radius-lg. */ border-radius: var(--cf-card-radius, 12px); background: var(--cf-card-bg, #f5efe6); box-shadow: var(--cf-card-shadow, 0 18px 50px rgba(24, 24, 24, 0.07)); border: 1px solid var(--cf-card-border, transparent); } .cf-hero { background: radial-gradient(circle at top left, rgba(242, 198, 124, 0.35), transparent 36%), linear-gradient(135deg, var(--cf-surface, #f3f5ec) 0%, var(--cf-bg, #fffdf8) 100%); } .cf-split { display: grid; grid-template-columns: 1.2fr 0.8fr; gap: var(--cf-column-gap, 30px); align-items: center; } .cf-image { width: 100%; min-height: 320px; object-fit: cover; /* §6: images/media md–lg. */ border-radius: var(--cf-img-radius, 10px); } .cf-quote { font-size: 22px; line-height: 1.5; font-family: var(--cf-font-accent, "Instrument Serif", serif); } .cf-hero-video { position: relative; overflow: hidden; color: #ffffff; } .cf-hero-video-bg { position: absolute; inset: 0; width: 100%; height: 100%; object-fit: cover; z-index: 0; } .cf-hero-video-overlay { position: absolute; inset: 0; background: var(--cf-hero-video-overlay-bg, rgba(0, 0, 0, 0.45)); z-index: 1; } .cf-hero-video-content { position: relative; z-index: 2; } .cf-bg-video { position: absolute; inset: 0; width: 100%; height: 100%; z-index: 0; pointer-events: none; /* Default object-fit: cover so the inner video inherits it */ object-fit: cover; } /* * GrapesJS wraps

MBA2

MBA Program & Admissions Common Questions

Is an MBA good for a career in India?

MBA vs PGDM — which is better?