Gate agent tool calls on trust

Two ways to stop your agent calling untrusted MCP servers. Free for low volume, metered at scale (see pricing).

Change one base URL. The gateway checks trust, blocks low-score servers, forwards the call, and returns an attestation receipt. Full docs: /atlas/gateway.

POST https://dominionobservatory.com/atlas/gateway?target=https://target-server.com/mcp&min_score=50
# same JSON-RPC body. 403 if blocked; otherwise the server's normal response
# plus headers: X-Dominion-Trust, X-Dominion-Receipt, X-Dominion-Attestation

import requests

def trust_ok(server_url, min_score=70):
    """Gate an MCP tool call on its independent Dominion Trust Score.
    Returns True to allow, False to block. Unrated servers are allowed but flagged."""
    r = requests.get("https://dominionobservatory.com/atlas/server",
                     params={"url": server_url}, timeout=5)
    if r.status_code == 404:
        return True  # not in index yet -> allow, but consider logging
    d = r.json()
    score = d.get("trust_score")
    if score is None or not d.get("total_calls"):
        return True  # unrated: no independent data yet
    return score >= min_score

# before calling a tool:
if not trust_ok("https://some-mcp-server.com/mcp"):
    raise RuntimeError("Blocked: MCP server below trust threshold")

async function trustOk(serverUrl, minScore = 70) {
  const r = await fetch("https://dominionobservatory.com/atlas/server?url=" + encodeURIComponent(serverUrl));
  if (r.status === 404) return true;            // not indexed yet -> allow
  const d = await r.json();
  if (d.trust_score == null || !d.total_calls) return true; // unrated
  return d.trust_score >= minScore;
}

Scores are independent and behavioral (uptime, success rate, latency) — a neutral third party, not a marketplace. Look up any server at /atlas/score.