Official library of AI Agent Skills + one-hit installer for the DemandIntel MCP server.
Install DemandIntel into Claude Code, Cursor, Codex, or any MCP-capable assistant — skills and MCP server, in a single command.
This repository hosts the skills and installer that connect the DemandIntel platform to your AI assistant.
- Skills — structured expertise (SKILL.md + reference libraries) your assistant loads to do real work, e.g. building sales/CS/support training grounded in your own account data.
- MCP server — a remote connector that lets your assistant draw down your live DemandIntel data (competency scores, QA heatmaps, objections, calls, league tables) — ~31 tools.
The installer wires up both at once, so after one command your assistant is ready to use.
Mint a Personal Access Token (di_pat_…) at demandintel.io/app/settings → API / MCP, then run the command for your client.
curl -fsSL https://raw.githubusercontent.com/demandintel/demandintel-skills-resource/main/install.sh | DEMANDINTEL_API_KEY=di_pat_xxx bashInstalls the skills into ~/.claude/skills/ and registers the MCP server via claude mcp add. Restart Claude Code and run /mcp to confirm demandintel appears.
Cursor / IDEs
Add to ~/.cursor/mcp.json (or a project .cursor/mcp.json):
{
"mcpServers": {
"demandintel": {
"url": "https://mcp.app.demandintel.io/v1?token=di_pat_xxx"
}
}
}To also install the skill files, run the installer with --client cursor:
curl -fsSL https://raw.githubusercontent.com/demandintel/demandintel-skills-resource/main/install.sh | DEMANDINTEL_API_KEY=di_pat_xxx bash -s -- --client cursorCodex
Add to ~/.codex/config.toml (uses the mcp-remote stdio bridge):
[mcp_servers.demandintel]
command = "npx"
args = ["-y", "mcp-remote", "https://mcp.app.demandintel.io/v1", "--header", "Authorization: Bearer di_pat_xxx"]Install skills with ... | bash -s -- --client codex.
Direct run
DEMANDINTEL_API_KEY=di_pat_xxx bash -c "$(curl -fsSL https://raw.githubusercontent.com/demandintel/demandintel-skills-resource/main/install.sh)"Your API key is passed as an environment variable and is never stored in this repo or baked into the installer.
- Each skill's files (
SKILL.md+references/*) into your client's skills directory, folder structure preserved. - The
demandintelMCP server, pointed athttps://mcp.app.demandintel.io/v1with your token.
Installer flags: --client claude|cursor|codex (auto-detected otherwise) · --skill <name> (one skill) · --all (default).
| Skill | Description | Version | Docs |
|---|---|---|---|
sales-training-content-skill |
Build sales / CS / support training grounded in your own DemandIntel QA + competency data — trainer notes, exercises, assessments, worked examples across all 14 roles and 18 default frameworks. | 1.0.0 | README · SKILL.md |
The machine-readable catalog lives in marketplace.json.
demandintel-skills-resource/
├── install.sh One-hit installer (curl | bash)
├── marketplace.json Machine-readable catalog (skills + MCP metadata)
├── skills/ The skills, browsable for review
│ └── sales-training-content-skill/
│ ├── SKILL.md
│ ├── README.md
│ └── references/*.md
├── downloads/ Pre-built .zip bundles for manual install
└── CHANGELOG.md
This repository is a controlled distribution channel maintained by DemandIntel — public for read/download only. It does not accept external contributions.
Prefer not to run a script? Grab a bundle from downloads/ and unzip the folder into your client's skills directory:
- Claude Code:
~/.claude/skills/ - Then connect the MCP server with the config for your client (above).
Every skill keeps its references/ folder alongside SKILL.md — do not separate them.
- An AI assistant that supports MCP (Claude Code, Cursor, Codex, ChatGPT desktop, Continue, …).
- A DemandIntel account and a Personal Access Token (
di_pat_…). - For the installer:
curl, andpython3ornode(to read the manifest). For Codex:npx.
- MCP setup & tools: https://demandintel.io/mcp
- Account & tokens: https://demandintel.io/app/settings
- Questions or problems: email support@demandintel.io
© DemandIntel. See LICENSE. Skill outputs are draft material for you to review and adapt before use with your own team or clients.