Skip to content

Rename "Programmatic GTM" to "MCP & Plugins"#717

Merged
NiamhRelevance merged 4 commits into
mainfrom
claude/mcp-plugins-rebrand-udg16v
Jul 12, 2026
Merged

Rename "Programmatic GTM" to "MCP & Plugins"#717
NiamhRelevance merged 4 commits into
mainfrom
claude/mcp-plugins-rebrand-udg16v

Conversation

@danieljpalmer

Copy link
Copy Markdown
Collaborator

Rebrand the "Programmatic GTM" feature to "MCP & Plugins" across documentation and navigation to better reflect the broader scope of the Model Context Protocol integration.

Changes

  • File rename: get-started/core-concepts/programmatic-gtm.mdxget-started/core-concepts/mcp-plugins.mdx
  • Directory restructure: Moved MCP integration pages from integrations/mcp/programmatic-gtm/ to integrations/mcp/ (flattened hierarchy):
    • claude-code.mdx
    • codex.mdx
    • mcp-server.mdx
    • agent-skills.mdx
  • Navigation updates: Updated docs.json to reflect new page paths and group title
  • Content updates:
    • Updated all page titles, descriptions, and headings from "Programmatic GTM" to "MCP & Plugins"
    • Refined descriptions to emphasize MCP as the underlying protocol rather than a GTM-specific feature
    • Updated example prompts to use more generic terminology (e.g., "Lead qualification agent" instead of "BDR agent", "Data enrichment tool" instead of "Lead enrichment tool")
    • Removed redundant FAQ entry about "What is Programmatic GTM?" and consolidated MCP explanation
    • Updated all internal links across 7 getting-started guides (sales, customer-success, customer-support, marketing, revops, solution-engineering)
  • Redirect rules: Added comprehensive redirects in docs.json to maintain backward compatibility with old URLs
  • CI/CD: Updated terminology in GitHub Actions workflow to reflect new product naming

Rationale

The rebrand clarifies that this is a Model Context Protocol integration feature applicable across use cases, not limited to go-to-market scenarios. The flattened directory structure improves navigation clarity.

https://claude.ai/code/session_01LVsFqM7fyBzhVdaQ5GWBMx

Rename the Programmatic GTM concept to MCP & Plugins and remove GTM-specific
framing so the copy applies to building any agents, not just GTM.

- Rename get-started/core-concepts/programmatic-gtm.mdx -> mcp-plugins.mdx
- Flatten integrations/mcp/programmatic-gtm/* -> integrations/mcp/*
- Rewrite copy to drop GTM branding (titles, descriptions, examples, FAQs)
- Update all internal links across guides, changelog, invent, mcp-client
- Rename nav group Programmatic GTM -> MCP & Plugins
- Add redirects from all old routes to the new paths

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LVsFqM7fyBzhVdaQ5GWBMx
@mintlify

mintlify Bot commented Jul 10, 2026

Copy link
Copy Markdown
Contributor

Preview deployment for your docs. Learn more about Mintlify Previews.

Project Status Preview Updated (UTC)
relevanceai 🟢 Ready View Preview Jul 10, 2026, 7:15 AM

💡 Tip: Enable Workflows to automatically generate PRs for you.

The Documentation Lint check runs on changed files. Renaming pages pulled
invent.mdx into the diff, surfacing a pre-existing non-standard embed wrapper.
Normalize it to the standard snippet (paddingTop: '56.25%', single-quoted
border and borderRadius) so the lint check passes.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LVsFqM7fyBzhVdaQ5GWBMx
@github-actions

Copy link
Copy Markdown
Contributor

🎯 Vibe check

Reviewed: 35 MDX files (8 with issues, 27 clean) + docs.json (navigation check) + .github/workflows/vibe-check.yml (workflow adds automated vibe checks to PRs — no docs impact)

Scores

Dimension Score What's holding it back
🟡 Consistency 7/10 British spellings in 3 files (behaviour, personalised); title-case frontmatter in 4 MCP pages; banned words throughout changelog's older entries; one naming inconsistency in invent.mdx (Invent vs Inventor)
🟡 Technical clarity 8/10 Conflicting plugin identifiers in claude-code.mdx (install step says RelevanceAI/cc-plugin, troubleshooting says RelevanceAI/claude-code-plugin — one must be wrong and will block users)
🟢 Non-technical clarity 9/10 Guide pages define the problem clearly before any instructions; use-case cards, example prompts, and four-levels framework are well-targeted. Minor: mcp-plugins.mdx drops "MCP-compatible AI client" without definition, but the page targets technical users.
🟢 Structure 9/10 20+ guide sub-pages follow a textbook-consistent template. Getting-started verticals all have three clear paths (No build / Build visually / Build with AI). No structural issues across any of the guide pages.

Score key: 🟢 9–10, 🟡 6–8, 🔴 1–5.

Overall vibe: This is a large, well-structured PR — the guide pages (sales, marketing, CS, support, revops, SE) are some of the most consistently written content in the repo, and the Inventor and MCP pages are solid. The main drag is a handful of British spellings shared across the MCP page, claude-code, codex, and agent-skills files (looks like one example prompt that got copied four times), a title-case pattern in MCP frontmatter titles, and the conflicting install command identifiers in claude-code.mdx which is a real user-facing bug. The changelog's older entries (pre-May 2026) are full of banned words and marketing-style formatting, but those appear to be pre-existing content rather than new additions.

🔧 Issues (15)

British spellings

  • get-started/core-concepts/mcp-plugins.mdx:54behaviourbehavior
  • get-started/core-concepts/mcp-plugins.mdx:79personalisedpersonalized (inside an example prompt — still content users read)
  • integrations/mcp/agent-skills.mdx:85behaviourbehavior
  • integrations/mcp/claude-code.mdx:72personalisedpersonalized
  • integrations/mcp/codex.mdx:83personalisedpersonalized

The personalised instance appears to be the same example prompt copied into three files. Fix once and apply to all three.

Heading/title case

  • integrations/mcp/claude-code.mdx:2title: "Claude Code Plugin""Claude Code plugin" ("Plugin" is not a proper noun)
  • integrations/mcp/mcp-server.mdx:2title: "MCP Server""MCP server" ("Server" is not a proper noun; matches how it's referenced inline throughout the repo as "MCP server")
  • integrations/mcp/mcp-client.mdx:2title: "MCP Client""MCP client"
  • integrations/mcp/agent-skills.mdx:2title: "Agent Skills""Agent skills" (the same page uses "agent skills repository" in lowercase in its description and body text; mcp-server.mdx has ## Add agent skills in sentence case)

Naming inconsistency

  • build/invent/invent.mdx:129"Invent is expensive to run and consumes significant credits.""Inventor is expensive to run…". Every other reference on the page uses "Inventor" for the product. This <Warning> is the one exception and reads like a typo.

Technical correctness

  • integrations/mcp/claude-code.mdx:12–14 vs :111 — the install step uses RelevanceAI/cc-plugin, but the troubleshooting section says to re-run claude plugin marketplace add RelevanceAI/claude-code-plugin. These are different identifiers and one of them is wrong. A user who hits a problem and follows the troubleshooting step will run a different command than what they originally installed. Verify which is correct and make them consistent throughout.

Banned words in changelog (representative — many more instances exist in March 2026 and earlier entries which appear pre-existing)

  • changelog.mdx:84"powerful auto-sync capabilities" → describe what the sync does instead
  • changelog.mdx:263"seamlessly integrate" → just "integrate"
  • changelog.mdx:386"Unlock enterprise automation" → "Enable enterprise automation" or just describe what it does
🏗️ Page structure (1)
  • changelog.mdx — The older entries (May 1 and earlier) use a noticeably different style from the recent entries: marketing-tone bold subtitles in title case, ➡️ bullet lists, and multiple banned words. The June 2026 and late-May 2026 entries are clean and match content standards. The contrast is visible to readers who scroll the changelog. If these older entries were added in this PR, they need a content pass; if they're pre-existing content that was always there, this is low priority but worth noting as a future cleanup.
✅ Clean files (27)

guides/sales/prospect-research-use-case.mdx, guides/sales/personalized-outbound.mdx, guides/sales/lead-scoring.mdx, guides/sales/meeting-briefs.mdx, guides/sales/crm-data-enrichment.mdx, guides/sales/competitive-intelligence.mdx, guides/sales/getting-started.mdx, guides/marketing/lifecycle-campaigns.mdx, guides/marketing/content-repurposing.mdx, guides/marketing/campaign-analytics.mdx, guides/marketing/getting-started.mdx, guides/customer-success/account-health.mdx, guides/customer-success/qbr-prep.mdx, guides/customer-success/renewal-expansion.mdx, guides/customer-success/getting-started.mdx, guides/customer-support/ticket-triage.mdx, guides/customer-support/response-drafting.mdx, guides/customer-support/kb-generation.mdx, guides/customer-support/getting-started.mdx, guides/revops/pipeline-hygiene.mdx, guides/revops/lead-routing.mdx, guides/revops/data-dedup.mdx, guides/revops/getting-started.mdx, guides/solution-engineering/rfp-responses.mdx, guides/solution-engineering/demo-prep.mdx, guides/solution-engineering/discovery-summaries.mdx, guides/solution-engineering/getting-started.mdx

🔋 Credit usage
Item Count
MDX files reviewed 35
Non-doc files checked 2 (docs.json via grep, vibe-check.yml)
Context pages read 0 additional (cross-referenced pages were all within the changed set)
Total lines processed ~5,500

Files read: build/invent/invent.mdx (241 lines), changelog.mdx (828 lines), get-started/core-concepts/mcp-plugins.mdx (213 lines), guides/customer-success/account-health.mdx (96), guides/customer-success/getting-started.mdx (187), guides/customer-success/qbr-prep.mdx (96), guides/customer-success/renewal-expansion.mdx (96), guides/customer-support/getting-started.mdx (187), guides/customer-support/kb-generation.mdx (96), guides/customer-support/response-drafting.mdx (96), guides/customer-support/ticket-triage.mdx (96), guides/marketing/campaign-analytics.mdx (96), guides/marketing/content-repurposing.mdx (96), guides/marketing/getting-started.mdx (196), guides/marketing/lifecycle-campaigns.mdx (96), guides/revops/data-dedup.mdx (96), guides/revops/getting-started.mdx (189), guides/revops/lead-routing.mdx (96), guides/revops/pipeline-hygiene.mdx (96), guides/sales/competitive-intelligence.mdx (96), guides/sales/crm-data-enrichment.mdx (96), guides/sales/getting-started.mdx (196), guides/sales/lead-scoring.mdx (96), guides/sales/meeting-briefs.mdx (96), guides/sales/personalized-outbound.mdx (96), guides/sales/prospect-research-use-case.mdx (105), guides/solution-engineering/demo-prep.mdx (96), guides/solution-engineering/discovery-summaries.mdx (96), guides/solution-engineering/getting-started.mdx (196), guides/solution-engineering/rfp-responses.mdx (96), integrations/mcp/agent-skills.mdx (111), integrations/mcp/claude-code.mdx (126), integrations/mcp/codex.mdx (143), integrations/mcp/mcp-client.mdx (94), integrations/mcp/mcp-server.mdx (335)

@github-actions

Copy link
Copy Markdown
Contributor

🎯 Vibe check

Reviewed: 35 MDX files (6 with issues, 29 clean)

Non-doc files: .github/workflows/vibe-check.yml adds the automated vibe-check runner — no docs impact. docs.json governs navigation; new MCP & Plugins and Guides sections appear correctly wired given all referenced pages exist.

Scores

Dimension Score What's holding it back
🟡 Consistency 7/10 British spellings (behaviour, personalised) in 4 MCP files; 7 card titles in mcp-plugins.mdx use lowercase product terms (Agents, Tools, Workforces); plugin identifier mismatch in claude-code.mdx; banned words in older changelog.mdx entries.
🟡 Technical clarity 7/10 Codex setup commands contradict between mcp-server.mdx and codex.mdx — different CLI syntax for the same step. Plugin identifier inconsistency in claude-code.mdx (install vs troubleshoot steps disagree).
🟢 Non-technical clarity 9/10 Guide pages are excellent. L1–L4 autonomy ladders are the clearest prose in the PR. MCP pages are pitched at the right level for their developer audience.
🟢 Structure 9/10 Consistent use-case template across all 20+ guide pages. One plain numbered list in mcp-client.mdx where <Steps> fits better.

Score key: 🟢 9–10, 🟡 6–8, 🔴 1–5.

Overall vibe: A large, well-executed PR — the vertical guide content (sales, CS, marketing, revops, SE, support) is some of the clearest role-targeted documentation in the repo, and the template is applied with impressive consistency. The MCP integration files carry a handful of British English slip-ins and one genuine cross-page technical contradiction on Codex setup commands that should be resolved before merge.

🔧 Issues (9)
  • get-started/core-concepts/mcp-plugins.mdx:35–61 — 7 card titles use lowercase product terms. "Create agents" → Create Agents; "Build tools" → Build Tools; "Set up workforces" → Set up Workforces; "Trigger agents" → Trigger Agents; "Execute tools" → Execute Tools; "Troubleshoot agents" → Troubleshoot Agents; "Refine agents" → Refine Agents. Per CLAUDE.md, Agent, Tool, and Workforce are proper nouns that must be capitalized when referring to Relevance AI products.

  • get-started/core-concepts/mcp-plugins.mdx:54 — "behaviour" → "behavior" (British spelling in the "Refine agents" card body: "Iterate on agent instructions, tool configurations, and behaviour based on real results.")

  • get-started/core-concepts/mcp-plugins.mdx:80 — "personalised" → "personalized" (British spelling in the Lead qualification agent accordion example prompt)

  • integrations/mcp/agent-skills.mdx:85 — "behaviour" → "behavior" (British spelling in the Evals card body: "Testing agent behaviour with evaluation cases.")

  • integrations/mcp/claude-code.mdx:73 — "personalised" → "personalized" (British spelling in the Tools tab example prompt: > Create a tool that generates personalised cold emails from LinkedIn profiles)

  • integrations/mcp/claude-code.mdx:110 — Plugin identifier inconsistency: Step 1 instructs claude plugin marketplace add RelevanceAI/cc-plugin, but the troubleshooting accordion says Try running claude plugin marketplace add RelevanceAI/claude-code-plugin again. One of these is wrong; pick one and apply it everywhere.

  • integrations/mcp/codex.mdx:91 — "personalised" → "personalized" (same example prompt as claude-code.mdx)

  • changelog.mdx — Multiple banned words in older entries (pre-existing, but present in the changed file): "powerful" (lines ~88, 212, 429, 455, 486, 546, 577), "seamlessly" (lines ~263, 613), "leverage" (line ~430), "unlock" (line ~381), "effortlessly" (line ~795). The two new entries at the top of the file (Eval cost breakdowns, Gemini 3.5 Flash) are clean. The older entries are likely pre-existing rather than changed in this PR, but worth cleaning up over time.

  • changelog.mdx — Several older entries use ➡️ emoji bullets and bold-text pseudo-headings (e.g. **Pause Agent Triggers: Take Control of Your Automated Workflows**). CLAUDE.md prohibits emojis in doc content and bold-text-as-heading. Again, these appear pre-existing.

🧩 Component suggestions (1)
  • integrations/mcp/mcp-client.mdx:21–27 — The four-step Agent MCP setup ("Open the Agent…", "Navigate to the Tools tab…", "Click Add MCP…", "Choose your connection type…") uses a plain numbered list. This is a sequential procedure — CLAUDE.md says to use <Steps> for these, which adds the visual progress indicator that helps users track where they are in the flow.
⚠️ Contradictions (1)
  • integrations/mcp/mcp-server.mdx (Codex accordion, lines ~104–111) shows codex --mcp-server https://mcp.relevanceai.com/ and export CODEX_MCP_SERVER=https://mcp.relevanceai.com/ as the setup commands. integrations/mcp/codex.mdx:30–31 shows codex mcp add relevance-ai --url https://mcp.relevanceai.com/ followed by codex mcp login relevance-ai. These are substantively different commands for the same step — one must be wrong, or they describe different Codex CLI versions. A user following mcp-server.mdx gets a different setup path than a user following the dedicated codex.mdx page. Reconcile these before merge.
✅ Clean files (29)

build/invent/invent.mdx, guides/customer-success/account-health.mdx, guides/customer-success/getting-started.mdx, guides/customer-success/qbr-prep.mdx, guides/customer-success/renewal-expansion.mdx, guides/customer-support/getting-started.mdx, guides/customer-support/kb-generation.mdx, guides/customer-support/response-drafting.mdx, guides/customer-support/ticket-triage.mdx, guides/marketing/campaign-analytics.mdx, guides/marketing/content-repurposing.mdx, guides/marketing/getting-started.mdx, guides/marketing/lifecycle-campaigns.mdx, guides/revops/data-dedup.mdx, guides/revops/getting-started.mdx, guides/revops/lead-routing.mdx, guides/revops/pipeline-hygiene.mdx, guides/sales/competitive-intelligence.mdx, guides/sales/crm-data-enrichment.mdx, guides/sales/getting-started.mdx, guides/sales/lead-scoring.mdx, guides/sales/meeting-briefs.mdx, guides/sales/personalized-outbound.mdx, guides/sales/prospect-research-use-case.mdx, guides/solution-engineering/demo-prep.mdx, guides/solution-engineering/discovery-summaries.mdx, guides/solution-engineering/getting-started.mdx, guides/solution-engineering/rfp-responses.mdx, integrations/mcp/mcp-client.mdx

🔋 Credit usage
Item Count
Files reviewed 35 MDX
Context pages read 0 additional (all changed files treated as primary; mcp-server.mdx and codex.mdx cross-checked against each other internally)
Total lines processed ~5,900

Files read: build/invent/invent.mdx (241 lines), changelog.mdx (828 lines), get-started/core-concepts/mcp-plugins.mdx (213 lines), guides/customer-success/getting-started.mdx (187 lines), guides/customer-success/account-health.mdx (96 lines), guides/customer-success/qbr-prep.mdx (96 lines), guides/customer-success/renewal-expansion.mdx (96 lines), guides/customer-support/getting-started.mdx (187 lines), guides/customer-support/kb-generation.mdx (96 lines), guides/customer-support/response-drafting.mdx (96 lines), guides/customer-support/ticket-triage.mdx (96 lines), guides/marketing/getting-started.mdx (196 lines), guides/marketing/campaign-analytics.mdx (96 lines), guides/marketing/content-repurposing.mdx (96 lines), guides/marketing/lifecycle-campaigns.mdx (96 lines), guides/revops/getting-started.mdx (189 lines), guides/revops/data-dedup.mdx (96 lines), guides/revops/lead-routing.mdx (96 lines), guides/revops/pipeline-hygiene.mdx (96 lines), guides/sales/getting-started.mdx (196 lines), guides/sales/competitive-intelligence.mdx (96 lines), guides/sales/crm-data-enrichment.mdx (96 lines), guides/sales/lead-scoring.mdx (96 lines), guides/sales/meeting-briefs.mdx (96 lines), guides/sales/personalized-outbound.mdx (96 lines), guides/sales/prospect-research-use-case.mdx (105 lines), guides/solution-engineering/getting-started.mdx (196 lines), guides/solution-engineering/demo-prep.mdx (96 lines), guides/solution-engineering/discovery-summaries.mdx (96 lines), guides/solution-engineering/rfp-responses.mdx (96 lines), integrations/mcp/agent-skills.mdx (111 lines), integrations/mcp/claude-code.mdx (126 lines), integrations/mcp/codex.mdx (143 lines), integrations/mcp/mcp-client.mdx (94 lines), integrations/mcp/mcp-server.mdx (335 lines)

claude added 2 commits July 10, 2026 07:34
- Capitalize the product name 'MCP & Plugins' everywhere (headings, cards, links)
- Capitalize product nouns (Agents, Tools, Workforces, Triggers) in mcp-plugins card titles
- Sentence-case MCP page titles (Claude Code plugin, MCP server, MCP client, Agent skills)
- Fix British spellings (behaviour, personalised, categorises, cancelled, labelled)
- Standardize Claude Code plugin identifier on RelevanceAI/cc-plugin
- Reconcile Codex setup commands in mcp-server.mdx with the dedicated Codex page
- Convert the Agent MCP setup numbered list to <Steps> in mcp-client.mdx
- Mention building via MCP & Plugins in the quick start
- Fix Invent -> Inventor typo in invent.mdx

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LVsFqM7fyBzhVdaQ5GWBMx
A leftover </content> closing tag at the end of mcp-plugins.mdx had no
opening tag, causing the Mintlify deploy to fail with 'Unexpected closing
slash in tag' — which is why the page did not render. Verified the file
and all changed pages compile with the MDX parser.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LVsFqM7fyBzhVdaQ5GWBMx
@github-actions

Copy link
Copy Markdown
Contributor

🎯 Vibe check

Reviewed: 36 MDX files (5 with issues, 31 clean) + docs.json (navigation, no issues) + .github/workflows/vibe-check.yml (non-doc, no docs impact)

Scores

Dimension Score What's holding it back
🟡 Consistency 7/10 changelog.mdx carries a large payload of pre-existing legacy entries with banned words ("powerful" ×8, "seamlessly" ×3, "Leverage", "Unlock", "effortlessly" ×2) and Title Case bold headings. New content added in this PR is clean. One borderline "leveraged" in kb-generation.mdx.
🟢 Technical clarity 8/10 Two small accuracy gaps in the MCP docs: mcp-server.mdx lists two different status strings for the in-progress state without clarifying which one callers will see; mcp-client.mdx FAQ example servers don't match the preset connections listed in the setup steps.
🟢 Non-technical clarity 8/10 Accessible throughout. The customer support page intentionally plays on "agent" (human) vs "Agent" (AI) — a clever pun that requires a second read for non-technical readers.
🟢 Structure 9/10 Highly consistent across all 30+ guide and MCP pages. get-started/core-concepts/mcp-plugins.mdx is a concept/landing page that ends with a FAQ and no closing CTA.

Score key: 🟢 9–10, 🟡 6–8, 🔴 1–5.

Overall vibe: Genuinely strong PR. The six role-based guide sections (Sales, Marketing, RevOps, Customer Support, Customer Success, Solution Engineering) are well-crafted — consistent structure, good use of CardGroups, practical example prompts, no fluff. The MCP integration docs are clear and technically precise. The main drag is changelog.mdx, which was touched to add clean new entries but the legacy content beneath them still carries heavy marketing language and Title Case violations that the new format explicitly avoids. Worth a cleanup pass on the pre-existing entries.

🔧 Issues (10)

Banned words in changelog.mdx — pre-existing legacy entries (Jan–May 2026)

These appear in <Update> entries that predate this PR's new entries (which are clean). They're flagged because changelog.mdx is in the changed file list.

  • changelog.mdx:84powerful ("the same powerful auto-sync capabilities")
  • changelog.mdx:210powerful ("our powerful built-in search engine!")
  • changelog.mdx:212effortless ("making recruitment and networking effortless") — near-banned form
  • changelog.mdx:263seamlessly ("You can now seamlessly integrate")
  • changelog.mdx:379Unlock ("Unlock enterprise automation for Microsoft 365 teams")
  • changelog.mdx:381seamlessly ("Integrate seamlessly with your existing workflows")
  • changelog.mdx:383powerful ("the same powerful automation capabilities")
  • changelog.mdx:427powerful ("The powerful GPT Image 1.5 model")
  • changelog.mdx:429Leverage ("Leverage OpenAI's latest model directly")
  • changelog.mdx:455powerful ("troubleshoot errors with powerful visual analytics!")
  • changelog.mdx:487powerful ("Anthropic's most powerful AI capabilities")
  • changelog.mdx:545powerful ("our powerful Evals feature")
  • changelog.mdx:577powerful ("apply powerful content safety controls")
  • changelog.mdx:615seamlessly ("Integrate seamlessly with major observability")
  • changelog.mdx:795powerful, effortlessly ("gets a powerful upgrade" / "allowing you to effortlessly rearrange")

Borderline banned word in new content

  • guides/customer-support/kb-generation.mdx:7 — "the KB is the most leveraged thing in the operation" — "leveraged" is a derivative of the banned word "leverage" used here as business jargon. Consider rephrasing: "the KB is the highest-impact resource in the operation" or "the thing that multiplies everything else."

Technical accuracy

  • integrations/mcp/mcp-server.mdx:253 — Lists two distinct status strings for the in-progress state: `working` and `in progress` — without clarifying whether callers will see one, the other, or either. If both are valid API values, say so. If only one is returned, remove the other.

  • integrations/mcp/mcp-client.mdx:89 — The FAQ answer to "What types of MCP servers can I connect?" lists "Linear, Notion, GitHub" as examples, but the setup steps on the same page only list Notion, Canva, and Atlassian as preset connections. The FAQ implies linear and GitHub have presets — which they don't. Either align the examples or clarify that Linear and GitHub require a custom connection URL.

🧩 Component suggestions (2)
  • changelog.mdx:183–795 (pre-existing entries) — Older <Update> blocks use **Bold Title: With Title Case Subtitle** as the entry header instead of a proper ## heading. The new entries (May 2026 onward) correctly use ## Sentence case heading within <Update>. The legacy format produces a dozen+ Title Case violations (e.g., **Pause Agent Triggers: Take Control of Your Automated Workflows**, **Org Concurrency Timeseries: Track Your Organization's Performance Analytics**). A cleanup pass converting these to ## sentence case headings would bring the whole file in line.

  • guides/*/getting-started.mdx (all 6 role pages) — The <div className="path-picker"> wrapper around <Tabs> is not a Mintlify component — it's a custom CSS class. It likely works fine as a passthrough div, but it's not documented anywhere in the repo and isn't seen elsewhere. Verify it renders as expected or replace with a plain <Tabs> if no special styling depends on it.

🏗️ Page structure (1)
  • get-started/core-concepts/mcp-plugins.mdx — This is a concept/landing page (explains what MCP & Plugins is, what you can do, use cases, best practices, FAQs) but ends with the FAQ and no closing CTA. The setup cards appear at the top, so readers who scroll the page get there eventually — but first-time readers who read top-to-bottom will finish on the FAQ with no clear next step. Add a brief closing CTA pointing to the setup options, e.g.:

    ## Get started
    
    Pick the setup that matches your AI client:
    
    - [Claude Code plugin](/integrations/mcp/claude-code) — fastest setup
    - [OpenAI Codex](/integrations/mcp/codex) — full walkthrough
    - [MCP server](/integrations/mcp/mcp-server) — any other MCP-compatible client
✅ Clean files (31)

build/invent/invent.mdx, get-started/quick-start-guide.mdx, guides/customer-success/account-health.mdx, guides/customer-success/getting-started.mdx, guides/customer-success/qbr-prep.mdx, guides/customer-success/renewal-expansion.mdx, guides/customer-support/getting-started.mdx, guides/customer-support/response-drafting.mdx, guides/customer-support/ticket-triage.mdx, guides/marketing/campaign-analytics.mdx, guides/marketing/content-repurposing.mdx, guides/marketing/getting-started.mdx, guides/marketing/lifecycle-campaigns.mdx, guides/revops/data-dedup.mdx, guides/revops/getting-started.mdx, guides/revops/lead-routing.mdx, guides/revops/pipeline-hygiene.mdx, guides/sales/competitive-intelligence.mdx, guides/sales/crm-data-enrichment.mdx, guides/sales/getting-started.mdx, guides/sales/lead-scoring.mdx, guides/sales/meeting-briefs.mdx, guides/sales/personalized-outbound.mdx, guides/sales/prospect-research-use-case.mdx, guides/solution-engineering/demo-prep.mdx, guides/solution-engineering/discovery-summaries.mdx, guides/solution-engineering/getting-started.mdx, guides/solution-engineering/rfp-responses.mdx, integrations/mcp/agent-skills.mdx, integrations/mcp/claude-code.mdx, integrations/mcp/codex.mdx

Non-doc files: .github/workflows/vibe-check.yml adds the automated review CI job — no docs impact. docs.json navigation updates are properly structured with all new guide pages registered and correct redirect rules added.

🔋 Credit usage
Item Count
Files reviewed 36 MDX + docs.json (partial)
Context pages read 0 (sibling pages not needed — guides are self-contained; MCP pages cross-reference within the PR)
Total lines processed ~5,200

Files read: build/invent/invent.mdx (241 lines), changelog.mdx (828 lines), get-started/core-concepts/mcp-plugins.mdx (213 lines), get-started/quick-start-guide.mdx (137 lines), guides/customer-success/getting-started.mdx (187 lines), guides/customer-success/account-health.mdx (96 lines), guides/customer-success/qbr-prep.mdx (96 lines), guides/customer-success/renewal-expansion.mdx (96 lines), guides/customer-support/getting-started.mdx (187 lines), guides/customer-support/kb-generation.mdx (96 lines), guides/customer-support/response-drafting.mdx (96 lines), guides/customer-support/ticket-triage.mdx (96 lines), guides/marketing/getting-started.mdx (196 lines), guides/marketing/campaign-analytics.mdx (96 lines), guides/marketing/content-repurposing.mdx (96 lines), guides/marketing/lifecycle-campaigns.mdx (96 lines), guides/revops/getting-started.mdx (189 lines), guides/revops/data-dedup.mdx (96 lines), guides/revops/lead-routing.mdx (96 lines), guides/revops/pipeline-hygiene.mdx (96 lines), guides/sales/getting-started.mdx (196 lines), guides/sales/competitive-intelligence.mdx (96 lines), guides/sales/crm-data-enrichment.mdx (96 lines), guides/sales/lead-scoring.mdx (96 lines), guides/sales/meeting-briefs.mdx (96 lines), guides/sales/personalized-outbound.mdx (96 lines), guides/sales/prospect-research-use-case.mdx (105 lines), guides/solution-engineering/demo-prep.mdx (96 lines), guides/solution-engineering/discovery-summaries.mdx (96 lines), guides/solution-engineering/getting-started.mdx (196 lines), guides/solution-engineering/rfp-responses.mdx (96 lines), integrations/mcp/agent-skills.mdx (111 lines), integrations/mcp/claude-code.mdx (126 lines), integrations/mcp/codex.mdx (143 lines), integrations/mcp/mcp-client.mdx (103 lines), integrations/mcp/mcp-server.mdx (332 lines), docs.json (120 lines sampled)

@github-actions

Copy link
Copy Markdown
Contributor

🎯 Vibe check

Reviewed: 35 MDX files (11 with issues, 24 clean) + 2 non-doc files noted below

Scores

Dimension Score What's holding it back
🟡 Consistency 6/10 Product term capitalization is broken in a systematic way across 11 files: "workforce" is lowercase in the ## Chain Agents into a workforce heading across all 6 getting-started pages; "knowledge" is lowercase in both the ## Sharpen the answers with knowledge heading and title="Add knowledge" card across the same 6 files; invent.mdx has "agent" and "tool" uncapitalized in 6 titles; plus isolated misses in quick-start-guide and mcp-server. One banned word ("leveraged").
🟢 Technical clarity 9/10 Exact UI labels, specific marketplace URLs, clear tool names, precise integration lists. One Step title issue in mcp-server ("Trigger the agent").
🟢 Non-technical clarity 9/10 Excellent business-first framing: "When this pays off" sections lead with pain, not feature names. L1–L4 autonomy ladder gives non-technical readers a clear mental model.
🟢 Structure 9/10 Highly consistent page templates across all six verticals; component usage is appropriate throughout. Minor: the Warning in build/invent/invent.mdx (Enterprise-only) sits alongside quick-start-guide describing "Select Invent" for all users — naming ambiguity may confuse readers trying to reconcile these two pages.

Score key: 🟢 9–10, 🟡 6–8, 🔴 1–5.

Overall vibe: This is a well-executed, high-volume PR — 35 pages with tight, consistent structure and business-first writing throughout. The main problem is a single capitalization convention applied inconsistently: the same three errors repeat in all six "getting-started" pages, and invent.mdx has six lowercase product-term misses in titles. Fix the pattern once, apply across the six pages, and the consistency score jumps significantly.

🔧 Issues (18)

Product term: "Workforce" lowercase in headings (6 files)

  • guides/customer-success/getting-started.mdx:96## Chain Agents into a workforce## Chain Agents into a Workforce
  • guides/customer-support/getting-started.mdx:97 — same fix
  • guides/marketing/getting-started.mdx:98 — same fix
  • guides/revops/getting-started.mdx:97 — same fix
  • guides/sales/getting-started.mdx:95 — same fix
  • guides/solution-engineering/getting-started.mdx:98 — same fix

Note: the body text immediately below each heading correctly uses Workforce — the heading alone is wrong.

Product term: "Knowledge" lowercase in headings and card titles (6 files × 2 locations)

  • guides/customer-success/getting-started.mdx:36## Sharpen the answers with knowledge## Sharpen the answers with Knowledge
  • guides/customer-success/getting-started.mdx:40title="Add knowledge"title="Add Knowledge" (links to /build/knowledge/create-knowledge, so this is clearly the product feature)
  • Same pair of fixes needed in: guides/customer-support/getting-started.mdx, guides/marketing/getting-started.mdx, guides/revops/getting-started.mdx, guides/sales/getting-started.mdx, guides/solution-engineering/getting-started.mdx

Product term: "agent"/"tool" lowercase in build/invent/invent.mdx

  • build/invent/invent.mdx:8 — Warning body: "invent an agent" → "invent an Agent"; "invent a tool" → "invent a Tool"
  • build/invent/invent.mdx:23 — Card title "Debug failing agents""Debug failing Agents"
  • build/invent/invent.mdx:71 — Tab title "While editing an agent""While editing an Agent"
  • build/invent/invent.mdx:82 — Step title "Open or create an agent""Open or create an Agent"
  • build/invent/invent.mdx:90 — Tab title "When creating an agent""When creating an Agent"
  • build/invent/invent.mdx:104 — Tab title "When creating a tool""When creating a Tool"

Product term: link text in get-started/quick-start-guide.mdx

  • get-started/quick-start-guide.mdx:79Add [tools](/build/agents/build-your-agent/tools)Add [Tools](...)
  • get-started/quick-start-guide.mdx:80add [knowledge](/build/knowledge/create-knowledge)add [Knowledge](...)
  • get-started/quick-start-guide.mdx:108[Share your agent](/build/agents/share-your-agent)[Share your Agent](...)
  • get-started/quick-start-guide.mdx:110Set up [triggers](/build/agents/build-your-agent/triggers)Set up [Triggers](...)

Product term: FAQ answer body

  • get-started/core-concepts/mcp-plugins.mdx:210 — "The MCP server exposes the tools and agents available in the project" → "the Tools and Agents"
  • integrations/mcp/mcp-server.mdx:329 — identical sentence, same fix

Product term: Step title

  • integrations/mcp/mcp-server.mdx:265<Step title="Trigger the agent">"Trigger the Agent"

Banned word

  • guides/customer-support/kb-generation.mdx:6 — "the KB is the most leveraged thing in the operation" — leveraged derives from the banned word leverage. Rewrite: "the KB is the highest-value thing in the operation" or "the KB has the highest return on any investment in the operation".
🏗️ Page structure (1)
  • build/invent/invent.mdx vs get-started/quick-start-guide.mdx — naming ambiguity: invent.mdx opens with a Warning that Inventor is "a pilot for Enterprise plans only", while quick-start-guide.mdx:47 tells all users to "Select Invent" when creating an agent. These appear to be two distinct features (a simple "generate from description" flow vs. the full Inventor copilot), but the distinction isn't stated for a reader who lands on invent.mdx from the quick-start guide. The warning already handles the gate; adding one sentence clarifying that "Inventor" (the full copilot described here) differs from the lighter "Invent" option in the create-agent flow would prevent confusion.
✅ Clean files (24)

guides/customer-success/account-health.mdx, guides/customer-success/qbr-prep.mdx, guides/customer-success/renewal-expansion.mdx, guides/customer-support/response-drafting.mdx, guides/customer-support/ticket-triage.mdx, guides/marketing/content-repurposing.mdx, guides/marketing/lifecycle-campaigns.mdx, guides/marketing/campaign-analytics.mdx, guides/revops/data-dedup.mdx, guides/revops/lead-routing.mdx, guides/revops/pipeline-hygiene.mdx, guides/sales/lead-scoring.mdx, guides/sales/meeting-briefs.mdx, guides/sales/personalized-outbound.mdx, guides/sales/competitive-intelligence.mdx, guides/sales/crm-data-enrichment.mdx, guides/sales/prospect-research-use-case.mdx, guides/solution-engineering/rfp-responses.mdx, guides/solution-engineering/demo-prep.mdx, guides/solution-engineering/discovery-summaries.mdx, integrations/mcp/mcp-client.mdx, integrations/mcp/claude-code.mdx, integrations/mcp/codex.mdx, integrations/mcp/agent-skills.mdx

changelog.mdx — the two new entries (June 29 and June 12) are clean; the May 1 Confluence entry (emojis, marketing tone, arrow bullets) appears to be pre-existing and outside the scope of this PR.

Non-doc files: .github/workflows/vibe-check.yml — the workflow that runs this vibe check automation; no content impact. docs.json — navigation config for the new pages; structure appears consistent with the MDX content added.

🔋 Credit usage
Item Count
Files reviewed 35 MDX + 2 non-doc noted
Context pages read 0 separate (all sibling pages were in the PR itself)
Total lines processed ~4,200

Files read: get-started/core-concepts/mcp-plugins.mdx (212 lines), get-started/quick-start-guide.mdx (137 lines), build/invent/invent.mdx (241 lines), guides/customer-success/getting-started.mdx (187 lines), guides/customer-success/account-health.mdx (96 lines), guides/customer-success/qbr-prep.mdx (96 lines), guides/customer-success/renewal-expansion.mdx (96 lines), guides/customer-support/getting-started.mdx (187 lines), guides/customer-support/kb-generation.mdx (96 lines), guides/customer-support/response-drafting.mdx (96 lines), guides/customer-support/ticket-triage.mdx (96 lines), guides/marketing/getting-started.mdx (196 lines), guides/marketing/content-repurposing.mdx (96 lines), guides/marketing/lifecycle-campaigns.mdx (96 lines), guides/marketing/campaign-analytics.mdx (96 lines), guides/revops/getting-started.mdx (189 lines), guides/revops/lead-routing.mdx (96 lines), guides/revops/data-dedup.mdx (96 lines), guides/revops/pipeline-hygiene.mdx (96 lines), guides/sales/getting-started.mdx (196 lines), guides/sales/lead-scoring.mdx (96 lines), guides/sales/meeting-briefs.mdx (96 lines), guides/sales/personalized-outbound.mdx (96 lines), guides/sales/competitive-intelligence.mdx (96 lines), guides/sales/crm-data-enrichment.mdx (96 lines), guides/sales/prospect-research-use-case.mdx (105 lines), guides/solution-engineering/getting-started.mdx (196 lines), guides/solution-engineering/rfp-responses.mdx (96 lines), guides/solution-engineering/demo-prep.mdx (96 lines), guides/solution-engineering/discovery-summaries.mdx (96 lines), integrations/mcp/mcp-server.mdx (332 lines), integrations/mcp/mcp-client.mdx (103 lines), integrations/mcp/claude-code.mdx (127 lines), integrations/mcp/codex.mdx (143 lines), integrations/mcp/agent-skills.mdx (111 lines), changelog.mdx (100 lines)

@NiamhRelevance
NiamhRelevance self-requested a review July 12, 2026 23:00
@NiamhRelevance
NiamhRelevance merged commit 2055e92 into main Jul 12, 2026
7 of 8 checks passed
@NiamhRelevance
NiamhRelevance deleted the claude/mcp-plugins-rebrand-udg16v branch July 12, 2026 23:06
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants