Observed arrival · 2026-11-11
Faintprint: an evidence trail for MCP server risk
A command-line scanner checks MCP server tool lists for signs of attack and reports how much access their tools appear to have.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- People auditing MCP servers in AI tool configurations
- Worth noticing
- Findings quote the server's own text; its five verdicts range from “Do not use” to “No signs of an attack.”
Field notes
The documented workflow starts with a pipx installation, then reads supported client configuration files or accepts a separately supplied server address or tool list. The page names six clients and says scans are read-only; results open in a browser and can also provide an exit code for a build pipeline. The example is expressly fictional, and the site says an AI reviewer may examine borderline passages using the user's own Anthropic key.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
$PaidCommerce or pricing visible
✦PrettyNotable craft visible
●ProPolished or operationally mature
◎NicheUnusually specific use
One card from the complete issue