One week inside Skillsync Five pieces in a week, and a brand their agents can keep running. See the week Scope a launch, see the work, install my skills and book a call, all without leaving your conversation.
TODO: the example chats below are illustrative. Swap in real transcripts once people use it.
Customize → Connectors → Add custom connector. Name it Ryan Skid and paste the URL. Works on Claude.ai and Claude Desktop, including the Free plan (one custom connector).
Choose No sign in. Type “brief Ryan Skid” in any chat.
chatgpt.com/plugins → + → Add custom MCP server. Paste the URL and choose No authentication, then Create.
Pick Ryan Skid from the tools menu in a new chat and ask for a brief.
Needs developer mode, which depends on your plan. ChatGPT can’t send a key header, so members use their personal key URL.
grok.com/connectors → New Connector → Custom. Paste the URL.
Ask Grok to brief Ryan Skid.
Custom connectors may need a paid plan.
Intelligence → Connectors → Add Connector → Custom MCP. Paste the URL.
Choose no authentication, then ask Le Chat to brief Ryan Skid.
From any project. It adds Ryan Skid to your MCP servers.
Tell Claude Code “brief Ryan Skid” and point it at the repo or the deck.
Works in the Codex CLI, IDE extension and app (they share one config).
Tell Codex “brief Ryan Skid”.
Or paste the config into ~/.cursor/mcp.json (or .cursor/mcp.json in a project).
In Agent mode, ask Cursor to brief Ryan Skid.
Or run the command, or Command Palette → MCP: Add Server → HTTP and paste the URL.
Open Copilot Chat in agent mode and ask for a brief.
Paste it into ~/.gemini/config/mcp_config.json (or .agents/mcp_config.json in a workspace). Antigravity needs serverUrl, not url. Also covers the agy CLI, which replaced Gemini CLI.
Ask the agent to brief Ryan Skid.
Paste it into ~/.config/devin/mcp_config.json (older Windsurf builds: ~/.codeium/windsurf/mcp_config.json), then refresh MCP servers in Cascade.
Ask Cascade to brief Ryan Skid.
One command, or paste the config into ~/.config/amp/settings.json.
Ask Amp to brief Ryan Skid.
Merge it into your Zed settings.json (Zed → Settings → Open Settings).
In the agent panel, ask to brief Ryan Skid.
CLI command, or in the extension open MCP Servers → Configure and paste the config. type must be streamableHttp.
Ask Cline to brief Ryan Skid.
Kiro asks you to confirm. Or paste the config into ~/.kiro/settings/mcp.json.
Ask Kiro to brief Ryan Skid.
Merge it into ~/.config/opencode/opencode.json or opencode.json in a project. Keep oauth: false.
Ask OpenCode to brief Ryan Skid. Works with any model, DeepSeek and Qwen included.
Or goose configure → Add Extension → Remote (Streamable HTTP), or paste the config.
Ask Goose to brief Ryan Skid.
Qwen Code needs httpUrl (url means the older SSE transport).
Ask Qwen Code to brief Ryan Skid.
Nous Research’s Hermes Agent. Add it to ~/.hermes/config.yaml.
Ask Hermes to brief Ryan Skid. Works with any model, DeepSeek and Qwen included.
Grok calls the server from xAI’s side; no client to install.
Ask the model to brief Ryan Skid.
The docs don’t say whether authorization takes the “Bearer ” prefix; the server accepts the bare key there.
DeepSeek’s app and API have no MCP support of their own. Run DeepSeek as the model inside OpenCode, Hermes Agent, Codex or Claude Code, then connect Ryan Skid there.
Its tab has the exact config.
It scopes the work, pulls the closest case study and drafts the brief, without the back-and-forth.
We demo at YC Demo Day in four weeks. Here’s the deck.
That’s a Launch: a launch film plus cutdowns, $28,000 fixed, three to four weeks. Closest work is Skillsync’s launch film, shipped in a week.
Find Devs That ShipSkillsync · Launch film Install the Skillsync motion skills so our team can make the next explainer.
Here are skillsync-motion-principles, skillsync-text-animation and magma-gradients, ready to add to your Claude Code skills. This is the explainer they came from.
Find Your Best WorkSkillsync · Explainer Show me something you’ve made for a payments company.
Hyperswitch, Juspay’s open-source payments switch: the founder film and the Product Hunt launch.
The Open-Source Future of PaymentsJuspay · Hyperswitch · Founder film We’re launching our data product to broadcasters. What would you make?
A Launch built around the live product. Closest work is CBS RomoVision, which won a Sports Emmy in 2022.
CBS RomoVisionCBS Sports · Broadcast AR · Emmy Book a call for Thursday afternoon.
Thursday at 2:00 pm PT is open. Booked a 15-minute intro with Ryan and sent the invite.
Intro callRyan Skid · 15 minutes The Model Context Protocol is an open standard that lets AI assistants like Claude use outside tools. Mine gives Claude eleven: scope a launch, compare the offers, show the closest work, fetch skills, check my calendar and book a call.
Any assistant that can add a remote MCP server over HTTP. The switcher above has exact steps for Claude, ChatGPT, Grok, Le Chat, Claude Code, Codex, Cursor, VS Code, Antigravity, Windsurf, Amp, Zed, Cline, Kiro, OpenCode, Goose, Qwen Code, Hermes Agent and the xAI API. DeepSeek and Qwen models work through any of those agents.
Describe what you’re launching and get a recommended offer with the price and timeline, see the case study closest to it, fetch any skill from the library and book an intro call. The brief it writes lands in my inbox.
Yes. Ask for any skill from the /skills library by name and Claude fetches it, ready to add to your setup. Paid skills need a membership key.
No. There’s nothing to sign in to. Add the URL and start. Paid skills take a membership key: most assistants send it as a header, and ChatGPT and Grok use your personal key URL instead.
Yes. The MCP is free. The work it scopes is priced per offer, and the price is in the reply before you book anything.
Pick your assistant at the top of this page and follow the three steps. In Claude Code it’s one command.
Have a question? Email or book a call.
