Fact Checking the Vellum Plugin Hub Announcement
A skill helps your assistant with one task. A plugin gives it a whole function it can run on its own.
That distinction is the difference between an assistant that drafts an email when you ask and one that reads your inbox, replies in your voice, and sends it without you sitting there.
Vellum just opened up its Plugin Hub, the free and open sourced plugin ecosystem for AI assistants.
Plugins save so much money on tokens because the AI doesn't have to research and teach itself how to do new work when you ask it. It just uses the right plugin to know how to do the work immediately.
These are some Vellum plugins that I'm using to 10x my agent harness:
1️⃣ 𝗠𝗮𝗿𝗸𝗲𝘁𝗶𝗻𝗴 𝗘𝘅𝗽𝗲𝗿𝘁 turns your assistant into a full marketing department. Positioning, launches, content, SEO and GEO, competitor teardowns, board reporting, all on demand instead of you prompting for each piece separately.
2️⃣ 𝗔𝗜 𝗛𝗲𝗿𝗼 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿 𝗞𝗶𝘁 packages five engineering skills from Matt Pocock's playbook into one plugin, so your assistant plans, writes PRDs, breaks work into vertical slices, and runs TDD loops the way a senior engineer would.
3️⃣ 𝗗𝘆𝗻𝗮𝗺𝗶𝗰 𝗡𝗼𝘁𝗰𝗵 puts your assistant inside your MacBook's notch. Click to chat, hold Ctrl+Option to talk, watch it work through a floating task bulb with live replies streaming in.
Now, on top of that, you can build your own custom plug-ins for your use case without you don't need to be a developer to do this either. Vellum has a plugin builder skill built in, describe what you want your assistant to do and it generates the plugin for you.
The part that made me trust it enough to connect real accounts: credentials stay local.
So when a plugin needs your email, your banking, your code repos, that access lives on your machine and never reaches the model. Install a plugin on your Mac app and it carries over to every other place you use your assistant like Slack and Telegram with the same memory intact.
This post is mostly accurate, but it overstates the impact of plugins on token savings and makes a few claims that need nuance.
Fact check
"Plugins are a game changer and major token saver for everyone using AI."
🟡 Partly accurate.
Plugins can reduce prompt length by encapsulating reusable instructions and tools, which may lower token usage in some workflows. However, the amount saved depends on the plugin design and the task. There is no guarantee of "major" savings for every user. (Vellum)
"A skill helps your assistant with one task. A plugin gives it a whole function it can run on its own."
🟡 Mostly accurate.
Vellum defines a plugin as a bundle that can include skills, tools, hooks, and other capabilities, making it broader than an individual skill. However, skills aren't necessarily limited to a single task—they can contain extensive instructions, examples, and reusable knowledge. (Vellum)
"That distinction is the difference between an assistant that drafts an email... and one that reads your inbox, replies in your voice, and sends it."
🟡 Needs nuance.
Plugins can enable these capabilities, but actions like sending emails depend on the available integrations, permissions, and trust settings. Vellum generally requires approval for sensitive actions by default. (Vellum)
"Vellum just opened up its Plugin Hub, the free and open sourced plugin ecosystem for AI assistants."
✅ Accurate.
Vellum recently introduced its Plugin Hub, with open-source, reviewed plugins available for installation. (Vellum)
"Plugins save so much money on tokens because the AI doesn't have to research and teach itself how to do new work."
🟡 Oversimplified.
Plugins can reduce repeated instructions and provide reusable capabilities, but they do not eliminate reasoning or retrieval costs. Token savings vary widely depending on the workflow and cannot be generalized. (Vellum)
"Marketing Expert turns your assistant into a full marketing department."
🟡 Marketing language.
The Marketing Expert plugin exists and provides marketing-related capabilities, but describing it as a "full marketing department" is promotional rather than factual. (Vellum)
"AI Hero Engineer Kit packages five engineering skills..."
✅ Mostly accurate.
Vellum offers engineering-focused plugins that bundle multiple reusable capabilities together. The exact contents should be verified against the current plugin version since plugins can evolve. (Vellum)
"Dynamic Notch puts your assistant inside your MacBook's notch."
✅ Accurate.
Vellum includes a Dynamic Notch interface for its macOS desktop application. (Vellum)
"You can build your own custom plugins without being a developer."
✅ Accurate.
Vellum includes tools that help users create plugins through natural language rather than writing everything manually. (Vellum)
"Credentials stay local."
🟡 Needs nuance.
When self-hosting, credentials are stored locally in the macOS Keychain. In Vellum Cloud, credentials are stored in an isolated encrypted vault rather than on the user's local machine. So "stay local" is not universally true. (Vellum)
"That access lives on your machine and never reaches the model."
✅ Mostly accurate.
Vellum states that credentials are handled by a separate credential system and are never exposed to the AI model. The model receives tool results, not the raw secrets. (Vellum)
"Install a plugin on your Mac app and it carries over to Slack and Telegram with the same memory intact."
🟡 Needs nuance.
This is true when those channels are connected to the same Vellum assistant and workspace. It's not a property of the plugin alone; it depends on Vellum's shared assistant architecture. (Vellum)
Overall verdict
Accuracy: 8.5/10
The description of Vellum's Plugin Hub and security architecture is largely accurate. The weakest parts are the sweeping claims about token savings, the generalized statement that credentials always stay local, and several pieces of promotional wording that overstate what plugins can do.
Real vs Fluff
🟢 85% Real | 🟡 15% Fluff
Real (85%)
Vellum Plugin Hub exists.
Plugins bundle multiple capabilities beyond individual skills.
Custom plugin creation is supported.
Dynamic Notch interface is available.
Credentials are isolated from the AI model.
Plugins are open source and reviewed before listing.
Shared assistant memory works across supported channels.
Fluff / Needs correction (15%)
"Major token saver for everyone" isn't universally true.
Plugins don't eliminate model reasoning or research costs.
"Full marketing department" is marketing language.
"Credentials stay local" only applies to self-hosted deployments; cloud deployments use an encrypted credential vault.
Cross-platform memory depends on using the same Vellum assistant, not simply installing a plugin.