Plugins Turn AI From Assistant to Operator
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 partly accurate, but it mixes correct concepts with several marketing claims and a few technically inaccurate statements.
Fact check
"Plugins are a game changer and major token saver..."
🟡 Opinion.
Plugins can reduce prompts and context, but whether they are a "major token saver" depends on how they're implemented.
"A skill helps your assistant with one task. A plugin gives it a whole function it can run on its own."
🟡 Oversimplified.
There is no universal distinction between "skills" and "plugins." Different AI platforms define these terms differently. In some systems, a skill can also invoke tools and perform complex workflows.
"...reads your inbox, replies in your voice, and sends it..."
✅ Technically possible.
With appropriate permissions, plugins/tools can automate email workflows.
"Vellum just opened up its Plugin Hub, the free and open sourced plugin ecosystem for AI assistants."
🟡 Needs current verification.
The launch, "free," "open sourced," and current availability should be confirmed because these can change over time.
"Plugins save so much money on tokens because the AI doesn't have to research and teach itself..."
🟡 Partially true, but overstated.
Plugins can reduce context length and repeated prompting, but they do not inherently reduce token usage. Many plugins still require prompts, tool calls, and model reasoning.
The primary benefit is better capability and reliability, not guaranteed token savings.
Marketing Expert plugin...
✅ Likely accurate as a description of what the plugin is intended to do.
AI Hero Engineer Kit...
✅ Likely accurate if those capabilities are documented by the plugin.
Dynamic Notch...
✅ Plausible.
This sounds like a macOS interface plugin.
"Build your own plugins without being a developer."
✅ Reasonable.
No-code plugin builders are common.
"Credentials stay local."
🟡 Needs verification.
This depends entirely on Vellum's architecture. It's a significant security claim and should be supported by official documentation.
"...access lives on your machine and never reaches the model."
🟡 Needs nuance.
Many secure AI systems keep credentials local or server-side and avoid exposing them to the LLM, but whether this is always true for Vellum depends on implementation details.
"...same memory intact."
🟡 Marketing language.
Cross-platform memory synchronization is possible, but "same memory" simplifies what is usually state synchronization.
Overall verdict
Accuracy: 7.5/10
The overall message—that plugins extend AI assistants with reusable capabilities—is correct.
The weakest parts are:
The universal definition of "skills vs plugins."
Overstating token savings.
Security claims ("credentials stay local") without cited documentation.
Time-sensitive product launch claims.
Real vs Fluff
🟢 70% Real | 🟡 30% Fluff
Real (70%)
Plugins can extend AI assistants.
Email and automation workflows are possible.
No-code plugin creation is increasingly common.
Cross-platform assistant workflows exist.
Fluff (30%)
"Major token saver."
Universal "skill vs plugin" distinction.
"10x my agent harness."
Security claims without verification.
"Never reaches the model."
Marketing language around memory and automation.