Which AI agents does Skillry support and how do I connect them?

Skillry supports Claude Code, Codex, and Cursor as its named agents, and its homepage shows a "Works with + 6 And more supported agents" indicator, meaning additional agents beyond the three named ones are also supported. To connect, you use the "Connect your agent" entry point on the site, then install a skill so it becomes callable inside your agent. The exact connection steps aren't spelled out in the available page content, so treat the flow below as the general pattern rather than a verified click-by-click path.

Agents Skillry names

The site's own title and description list three agents explicitly:

So if you use Claude Code, Codex, or Cursor, you're in the explicitly supported set. If you use something else, the "+ 6" line suggests it may still work, but the page doesn't name those agents — check the site directly before assuming your agent is covered.

How the connection model works

Skillry is a curated library of Agent Skills — reusable SKILL.md workflows. The description states these cover slides, web pages, images, and video. The key idea:

• A skill is a packaged workflow file (SKILL.md), not a plugin you configure per-agent from scratch.
• You install a skill from the library, and it becomes available to run inside your connected agent.
• The homepage frames this as "Install one, run it end to end, keep what it makes" — meaning the skill runs a full task and produces an output you retain.

This is why the same skill listing shows a "+ 6" badge next to each item: one skill is intended to work across the supported agents rather than being locked to a single one.

Connecting your agent

The homepage provides a "Connect your agent" entry point. Based on how skill libraries of this type generally work, the flow is:

1. Open the "Connect your agent" entry on Skillry.
2. Select or authorize your agent (Claude Code, Codex, Cursor, or another supported one).
3. Install a skill from the library — free skills are marked "Free" and install without a card or trial timer.
4. Invoke the skill inside your agent to run the workflow end to end.

Expected result: the installed skill is callable from within your agent, and running it produces the artifact the skill is designed for (a deck, a landing page, an image, a video, etc.).

The page content does not document the precise menu labels, auth method, or config file changes for each agent, so verify those on the site rather than relying on a guessed sequence.

Free vs. paid skills when connecting

The library separates Free Agent Skills from the rest. Free skills are described as installable with "No card, no trial timer." Examples shown include GrokBot Avatar (Image), Velocity-Matched UI Sting (Video), Editable Visual Deck (Slides), and Activity Calendar Chart (Web).

If you're just testing whether your agent connects and runs a skill, start with a free one — it removes the payment variable from the setup. Pricing for non-free skills lives at the site's Pricing page; the available content doesn't state amounts, so check there.

Practical notes

• Start with a free skill to confirm your agent is connected and can execute a workflow before spending anything.
• Match the skill to your output type — the library is organized by category (Web, Image, Slides, Video), so pick within the category you need.
• The "+ 6" is a claim, not a list. If your agent isn't Claude Code, Codex, or Cursor, confirm support on the site before building a workflow around it.
• One skill, multiple agents. The repeated "+ 6" badges suggest skills are meant to be portable across supported agents, so a skill you install for one agent may also run in another.