From c4f0db5719e629b5c0e9b45f056c2c5cf2ba0d00 Mon Sep 17 00:00:00 2001 From: Marco Sadjadi Date: Wed, 8 Jul 2026 23:13:53 +0200 Subject: [PATCH] =?UTF-8?q?fix(review):=20loop-2=20findings=20=E2=80=94=20?= =?UTF-8?q?login=20fallback,=20proof-band=20confidence,=20contrast,=20cont?= =?UTF-8?q?ent=20hedges?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Code review: login no longer strands visitors with zero sign-in methods when the providers fetch fails (falls back to SMS + error notice); skeleton while providers load instead of a blank card. Design re-audit: proof band flipped from apology to flex ('Don't take our word for it / Every claim links to its proof') and promoted to the primary heading tier; unverifiable '60 seconds'/'in minutes' speed claims dropped; fg-subtle body text bumped to fg-muted (AA contrast); marketplace section differentiated via border-y + elevated bg (adjacent hairlines removed); preview frame traffic lights on-palette; header h-12 -> h-14 (hero svh calc adjusted); emerald-400 -> --color-success token. Content QA: rest-api article description drift between page and registry resolved; ChatGPT plan-gating table and Atlassian Rovo SSE-cutoff claims now carry dated hedges. Co-Authored-By: Claude Fable 5 Claude-Session: https://claude.ai/code/session_01SXUwmPVRTD8AKQtio6gCN5 --- .../guides/chatgpt-mcp-connector/page.tsx | 4 +- .../guides/mcp-transports-explained/page.tsx | 3 +- apps/web/app/(marketing)/layout.tsx | 10 ++-- apps/web/app/(marketing)/page.tsx | 48 ++++++++++--------- apps/web/app/login/page.tsx | 22 ++++++++- apps/web/lib/articles.ts | 2 +- 6 files changed, 57 insertions(+), 32 deletions(-) diff --git a/apps/web/app/(marketing)/guides/chatgpt-mcp-connector/page.tsx b/apps/web/app/(marketing)/guides/chatgpt-mcp-connector/page.tsx index 18a06ba..477f33c 100644 --- a/apps/web/app/(marketing)/guides/chatgpt-mcp-connector/page.tsx +++ b/apps/web/app/(marketing)/guides/chatgpt-mcp-connector/page.tsx @@ -38,7 +38,9 @@ export default function Page() {

The plan limits, first — because they decide everything

Before writing a single prompt or line of code, check what your ChatGPT plan allows. - OpenAI gates custom MCP connectors by workspace type: + As of mid-2026, OpenAI gates custom MCP connectors by workspace type — verify against + OpenAI's current help docs before committing to a plan, since these limits have + shifted before:

diff --git a/apps/web/app/(marketing)/guides/mcp-transports-explained/page.tsx b/apps/web/app/(marketing)/guides/mcp-transports-explained/page.tsx index 156a76b..925e655 100644 --- a/apps/web/app/(marketing)/guides/mcp-transports-explained/page.tsx +++ b/apps/web/app/(marketing)/guides/mcp-transports-explained/page.tsx @@ -129,7 +129,8 @@ export default function Page() { Spec revision 2025-03-26 replaced HTTP+SSE with Streamable HTTP. The old transport still works where both sides keep supporting it, but the direction is one-way: platforms have been announcing removal dates through 2026 — Atlassian's Rovo MCP server, - for example, set its HTTP+SSE cutoff for June 30, 2026. New servers should not ship it. + for example, announced an HTTP+SSE cutoff of June 30, 2026. New servers should not ship + it.

Streamable HTTP: the current standard

diff --git a/apps/web/app/(marketing)/layout.tsx b/apps/web/app/(marketing)/layout.tsx index 6746330..24edca6 100644 --- a/apps/web/app/(marketing)/layout.tsx +++ b/apps/web/app/(marketing)/layout.tsx @@ -8,7 +8,7 @@ export default function MarketingLayout({ children }: { children: React.ReactNod return (
-
+