Files
Mattsson a80ce54b78 fix(mcp): eager-auth flag on the Grok connector links so Grok starts OAuth (#2167)
* fix(mcp): eager-auth flag on the Grok connector links so Grok starts OAuth

Live test after #2158: pasting the Grok URL into grok.com's custom
connector dialog listed all 150+ tools and never opened the sign-in. Grok
probes the URL without credentials, like claude.ai, and reads the lazy
200 on initialize as an authless server; only the 401 challenge starts
OAuth (#2159 fixed the same thing for the claude.ai link).

- lib/onboarding/checklist.ts: mcpServerUrl() builds the server URL with
  an optional eagerAuth flag; sideDoorServerUrl() gives the Grok side door
  auth=required and keeps ChatGPT lazy; claudeConnectorLink() reuses it.
  SIDE_DOORS / SideDoor move here from the component. Tests for all three.
- NewUserChecklist copies the door-specific URL (now with a client marker).
- ApiKeysPanel's Grok row copies the flagged URL, mirroring the Claude one.
- auth-mode.ts comment records the second consumer; registry entry's Grok
  step carries the flag; DECISIONS.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LhTJcwgzmN3TsLR8tVHwdi
Signed-off-by: Emil <emilmattsson14@gmail.com>

* docs(mcp): registry Claude.ai step carries auth=required too

Review pass on #2167: the registry entry flagged the Grok install URL
but left the Claude.ai step on the bare URL, which pre-fills "None" in
claude.ai's dialog (#2159). Same file, same flag, now consistent.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LhTJcwgzmN3TsLR8tVHwdi
Signed-off-by: Emil <emilmattsson14@gmail.com>

---------

Signed-off-by: Emil <emilmattsson14@gmail.com>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
2026-09-02 16:55:23 +02:00
..

Community Registry

This directory is the source of truth for the registry shown at gnubok.se/community/registry: skills, MCP servers, workflows and apps built on Accounted. The website syncs its registry pages from here, so adding an entry is a normal pull request against this repo.

Structure

registry/
  entries/   one .mdx file per registry entry
  authors/   one .mdx file per author profile

Files starting with _ are templates and are ignored by validation and sync.

Adding an entry

  1. Copy entries/_template.mdx to entries/<your-slug>.mdx. The filename is the slug and the public URL: gnubok.se/community/registry/<your-slug>.
  2. If this is your first contribution, copy authors/_template.mdx to authors/<your-handle>.mdx. The author field in your entry must match an author file, so first-time contributors add both files in the same PR.
  3. Validate locally: npm run validate:registry (or npx tsx scripts/validate-registry.ts without installing everything).
  4. Open a PR. Commits need a DCO sign-off (git commit -s), same as the rest of the repo; see CONTRIBUTING.md.

A maintainer reviews the entry (does it work, does it describe itself honestly, is the content safe) and merges. After merge the website pulls the entry in with its registry sync; expect it live on the site within a few days of merging.

Entry frontmatter

Required:

Field Type Notes
title string Shown as the card and page title
description string Card subtitle, max 500 chars; put detail in the body
slug string Must equal the filename
kind skill | mcp | workflow | app Which shelf it goes on
author string Handle of a file in authors/
status live | beta | archived Be honest; beta is fine
lang sv | en Language of the entry body
personas list Any of founder, finance, byra, developer
publishedAt date YYYY-MM-DD
updatedAt date YYYY-MM-DD, bump when you edit

At least one of installCommand, downloadUrl, repoUrl, externalUrl is required so readers can actually get the thing. Optional extras: featured (maintainer-set), oauthScopes, gnubokTools (which MCP tools it touches), requiresWriteScope, sieCompatible, version, faq (list of {q, a}), related (list of slugs), ogImageEyebrow.

Body rules

The body after the frontmatter is plain Markdown (headings, lists, tables, links, fenced code blocks). The website renders bodies through MDX, which evaluates raw tags (lowercase HTML included), {...} expressions and import/export statements at build time, so the validator rejects all of them. Fenced code blocks and backtick inline code are fine; they are displayed, never executed. Need a literal <tag> or {value} in prose? Put it in backticks.

Write the body like documentation, not a landing page: what it does, what it needs (scopes, API keys), what it will not do, and one honest limitation beats three superlatives.

What gets accepted

  • It must exist and work against Accounted today (the live app or the MCP server).
  • It must describe its write behavior truthfully. Anything that writes to the ledger goes through staged operations that a human approves; entries that work around that will not be listed.
  • No fees are charged for listing, and your IP stays yours. Entries are documentation and are contributed under this repo's license; the thing the entry points to keeps whatever license you gave it.