Skip to content

lastseen score

by lastseen.dev · remote · no sign-in to list tools · v1.0.1 · published by dev.lastseen

Trust scores and dated liveness/drift facts for MCP servers, from the lastseen.dev notary register.

Changes · 3
CompareRegistry entry

3 tools: 3 change things.

Scanned Oct 21, 2026

No check matched anything in the tool definitions. No finding is not a safety guarantee. How the checks work.

claude mcp add --transport http score https://score.lastseen.dev/mcp

Tools

Open a tool for its reasons and full description. Descriptions are the publisher's own words, shown as they are.

3 tools

Changes things 3Claude Code: ask
  • dependency_riskWhen to call: before you connect to or recommend an MCP server, when you care whether its declared dependencies were observed dead or dormant in the lastseen.dev register.No findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.
    Description, as the server wrote it
    When to call: before you connect to or recommend an MCP server, when you care whether its declared dependencies were observed dead or dormant in the lastseen.dev register. Returns: dated facts, not a score: depends_on[], dead_dependency_count, dormant_dependency_count, n, provenance. Empty depends_on or zero counts means no dependency edges were on record for that id as of the response, not a guarantee of zero risk. Example: {"id":"io.github.06ketan/medium-ops"} id: MCP schema-server identifier whose dependency edges to inspect (e.g. io.github.owner/name).
  • schema_historyWhen to call: before you rely on an MCP server's tools being stable, e.g.No findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.
    Description, as the server wrote it
    When to call: before you rely on an MCP server's tools being stable, e.g. you are about to integrate it, pin tool names, or check whether the tool list changed under the same server id. Returns: dated facts, not a score: first_seen, last_observed, drift_events, n, and observations[] with observation_date, tools_hash, tool_count (may be null), changed. Use trust_score for the derived score and dependency_risk for dependency mortality counts. Example: {"id":"io.github.06ketan/medium-ops"} id: MCP schema-server identifier (e.g. io.github.owner/name). Same id space as trust_score.
  • trust_scoreWhen to call: before you connect to, install, wire into an agent, or recommend an MCP server, to check what the public record shows about whether it is still maintained.No findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.
    Description, as the server wrote it
    When to call: before you connect to, install, wire into an agent, or recommend an MCP server, to check what the public record shows about whether it is still maintained. Pass the MCP schema-server id. Returns: JSON with trust_score (0 to 100 or null), display (score · depth · provisional), depth/depth_max, per-dimension values (liveness, maintenance, adoption, spec_currency, dependency_health; null = unknown, never invented), effective_weights, resolution method/confidence, status (scored | unresolved | insufficient_data), methodology URL. Depth 2 answers mortality more than merit; cite display as-is. Not a prediction, not advice, not a ranking of alternatives. Example: {"id":"io.github.06ketan/medium-ops"} id: MCP schema-server identifier (e.g. io.github.owner/name or owner/name). Not an npm package name, Action slug, or Docker image.

This is not a safety rating, and no finding is not a safety guarantee. The groups and findings come from fixed checks of the tool definitions the server returned on the date shown, not from its code or what it does when called. How the checks work.

Install

Pinned to version 1.0.1. Values in angle brackets are yours to fill in.

claude mcp add --transport http score https://score.lastseen.dev/mcp

Run in a terminal.

Claude Code permissions

Allow the reads, ask before a change, deny the deletes. Paste into .claude/settings.json.

.claude/settings.json
{
  "permissions": {
    "allow": [],
    "ask": [
      "mcp__score__trust_score",
      "mcp__score__schema_history",
      "mcp__score__dependency_risk"
    ],
    "deny": []
  }
}

Tools that only read are allowed, tools that change things ask first, tools that delete or cancel are denied. It assumes the server is installed as score. Review it before you use it.