Ask an assistant for today's internet-wide cyber threat band, score, critical advisories and data-source health.
Security teams often want one quick answer before digging into their own dashboards: is the internet having a bad day? ICHI, the Internet Cyber Health Index, publishes a daily, source-transparent reading of internet-wide threat conditions, and the ICHI MCP server lets an assistant fetch it on request. It is a small Python stdio server that reads ICHI's free public API. No key or account is involved.
get_current_index returns the headline score, a rounded score, the band and its range, when it was evaluated, the model version, and per-dimension sub-scores with readable names and their contributions. Freshness details are included too: whether the headline is live, its age, and any stale or impaired inputs.get_recent_alerts pulls the latest items, 5 by default and up to 50, from ICHI's critical-advisory RSS feed, with title, link, publish time, category and description.get_source_health reports overall status and confidence, counts of live, stale, delayed and failed feeds by tier, and which sources fed the score.Every response carries "ok": true, or "ok": false with an error message on timeouts or bad upstream data. The server does not crash on a malformed reply.
This is an outside-in directional signal. It is not a risk score for your organisation, not a forecast, and not a replacement for internal telemetry. The band, running from NORMAL up to CRITICAL, is the meaningful part; movement within a band is noise. A reading of 65 sits in ELEVATED, which spans 50 to 69. The index is re-evaluated roughly every ten minutes, and the alerts feed is observational and carries no score of its own.
You need Python 3.10 or newer. The simplest route is uvx, installing straight from the project's git source; it is not on PyPI. Pip works the same way with a git URL, after which you run ichi-mcp. Claude Desktop sometimes cannot find uvx on its PATH, in which case use its absolute path.
| Variable | Default | Purpose |
|---|---|---|
ICHI_TIMEOUT |
10 | HTTP timeout in seconds |
ICHI_BASE_URL |
ichi.global | API base, for testing only |
The ICHI MCP server adds no data and does no scoring; it relays what ICHI publishes. There is no caching, so each call is one HTTP request. Dimension names are mapped from API keys using the published methodology, and unknown keys pass through untouched. If the upstream schema changes, missing fields come back as null rather than errors. It is maintained by and for the index's creator.
More integration servers for assistants sit in the Other Integrations category.
Check the linked publisher documentation and your connected client's tool list. Names and availability can vary with the server version and account permissions.
| Tool | What it does |
|---|---|
| get_current_index | Current score, band and range, evaluation time, model, per-dimension sub-scores and freshness info. |
| get_recent_alerts | Latest 1-50 items (default 5) from ICHI's critical-advisory feed. |
| get_source_health | Overall status and confidence, feed counts by tier, and failed or stale feeds. |
{
"mcpServers": {
"ichi": {
"command": "uvx",
"args": [
"--from",
"git+https://github.com/bernik-bot/ichi-mcp",
"ichi-mcp"
]
}
}
}Requires Python 3.10+ and uv. If uvx is not on Claude Desktop's PATH, use its absolute path (which uvx). Pin a release with git+https://github.com/bernik-bot/ichi-mcp@v0.1.0.
| Variable | Description | Required |
|---|---|---|
| ICHI_TIMEOUT | HTTP timeout in seconds. Default 10. | Optional |
| ICHI_BASE_URL | API base URL, for testing only. Default https://ichi.global. | Optional |
Run an Actual Budget ledger from your AI agent, with phone approvals for every change on paired machines.
An Odoo module that serves a read-only MCP endpoint at /mcp, scoped to each user's own access rights.
Scans a Windows screen into a numbered element table so an agent clicks and types by id instead of guessing pixels.
Open-source screen recorder and editor whose MCP server lets an agent record, edit and export a demo video.
Agent memory that keeps every correction with its history, plus saved working state so a crashed run can resume.
Manage a Linux VPS, Postgres and Odoo databases and GitLab deploys over SSH, with previews, allowlists and an audit trail.