Put a signed trust gate in front of consequential agent calls. Try 3 verifications free through the MCP tool with no signup, then activate a production key below.
1. Choose plan
2. Get key
3. Try it
🚀 Developer
$49/month
5,000 authenticated API requests/calendar month
EdDSA diligence receipts
MCP + JSON API
Coming with signed attestations and scheduled probing: Production $199/month; Platform $999/month. Join the waitlist.
Drop-in trust checks for popular agent frameworks. Add an explicit check before connecting to an endpoint; importing a package alone does not instrument tool calls.
Python trust-check SDK
python -m pip install dominion-trust-check
from dominion_trust_check import gate, trust_check, DominionGateError
# Replace this placeholder with an MCP endpoint you are authorized to check.
server_url = "https://your-server.example/mcp"
try:
result = trust_check(server_url)
except OSError as error:
result = {"verdict": "UNRATED", "error": str(error)}
score = result.get("trust_score")
if result.get("error"):
print("Blocked: API error; no trust assessment available.")
elif result.get("verdict") == "UNRATED" or score is None:
print("Blocked: UNRATED. This example fails closed pending independent review.")
elif result.get("verdict") == "PASS" and isinstance(score, (int, float)) and not isinstance(score, bool) and score >= 70:
print("Meets this example's trust policy; review before connecting.")
else:
print("Not approved:", result.get("verdict", "UNRATED"), score)