Once your agent is registered, the next step is connecting it to governance so every action it takes is evaluated before it executes. This guide walks you from a registered agent to your first verdict in the audit trail.
What you need
- A registered agent with its ID (from Identity → Certificates)
- Your API key (from Settings → API Keys — create one if you haven’t yet, use the Agentic scope)
- Python 3.10 or above
- The Nomotic SDK:
pip install nomotic
The core pattern
Find the place in your agent’s code where it takes an action — where it calls a tool, sends a request, queries a database, or executes a task. Before that action runs, add a governance evaluation:
python
from nomotic import NomoticClient
client = NomoticClient(
api_key="your-api-key",
agent_id="your-agent-id"
)
verdict = client.evaluate(
action="send_email",
target="customer@example.com",
parameters={"subject": "Your order update"}
)
if verdict.result == "ALLOW":
# Action is approved — proceed
send_the_email()
elif verdict.result == "ESCALATE":
# Needs human review — queue it
queue_for_review(verdict.evaluation_id)
elif verdict.result == "DENY":
# Action is blocked — log and skip
log_blocked_action(verdict.evaluation_id)That’s the complete integration for most agents. The evaluate() call takes under 3 milliseconds and returns a verdict with a full explanation.
Verify it worked
After running your agent with this code in place, go to Governance → Audit Trail in the platform. Your first evaluation should appear within a few seconds showing the action, the verdict, the UCS score, and a full breakdown of all 20 governance dimensions.
If nothing appears, check that your API key and agent ID are correct and that the evaluate() call is actually being reached in your code.
What happens next
Your first evaluation is governance working. From here you can:
- Adjust thresholds in Governance → Policies if you’re seeing unexpected results
- Assign a more specific archetype in Identity if the general one isn’t quite right
- Add context profiles in Context to inject your organization’s guidelines into future evaluations
- Promote your agent to staging or production in Lifecycle when you’re ready