Skip to content
Tentacular Tentacular

Getting Started from a Scaffold

This guide walks through the most common path for creating a tentacle: finding a matching scaffold, configuring its parameters, and deploying it. This is Lifecycle B — scaffold used as-is.

Prerequisites

Section titled “Prerequisites”
  • tntc CLI installed and configured
  • A Kubernetes cluster with the MCP server installed

When to Use This Path

Section titled “When to Use This Path”

Use Lifecycle B when:

  • A scaffold closely matches what you need
  • You only need to fill in configuration values (endpoints, schedules, thresholds)
  • The scaffold’s node structure and data flow are correct for your use case

If you need to change the scaffold’s structure (add nodes, restructure config, modify the DAG), see Creating Custom Scaffolds for Lifecycle C and D.

Steps

Section titled “Steps”

1. Search for Scaffolds

Section titled “1. Search for Scaffolds”

Start by searching for scaffolds that match your goal:

Terminal window
tntc scaffold search "uptime monitor"
SOURCE    NAME              CATEGORY     COMPLEXITY  DESCRIPTION
public    uptime-tracker    monitoring   moderate    Probe HTTP endpoints every 5 minutes...
public    uptime-prober     monitoring   simple      Simple HTTP probe with alerting...

2. Review the Scaffold

Section titled “2. Review the Scaffold”

Get detailed information about the best match:

Terminal window
tntc scaffold info uptime-tracker
Name:         uptime-tracker
Display Name: Uptime Tracker
Source:        public
Category:     monitoring
Complexity:   moderate
Version:      1.0


Parameters:
  endpoints (list, required):
    HTTP endpoints to probe for uptime monitoring.
  latency_threshold_ms (number, optional, default: 2000):
    Maximum acceptable response time in milliseconds before alerting
  probe_schedule (string, optional, default: "*/5 * * * *"):
    Cron expression for how often to probe endpoints


Files:
  scaffold.yaml, workflow.yaml, params.schema.yaml, .secrets.yaml.example
  nodes/probe-endpoints.ts, nodes/store-results.ts, nodes/alert-failures.ts

3. Create the Tentacle

Section titled “3. Create the Tentacle”

Initialize a new tentacle from the scaffold. Use --enclave to place it in an enclave from the start:

Terminal window
tntc scaffold init uptime-tracker acme-uptime --enclave marketing-automations --no-params

This creates ~/tentacles/marketing-automations/acme-uptime/ with all scaffold files plus a tentacle.yaml recording the scaffold provenance.

Without --enclave, the tentacle is created at ~/tentacles/acme-uptime/ and you’ll need to specify --enclave at deploy time. Providing it during init keeps the workspace layout consistent with the three-layer model.

4. Configure Parameters

Section titled “4. Configure Parameters”

Read params.schema.yaml to understand what needs configuring. Each parameter has a path field telling you where in workflow.yaml the value lives.

Edit workflow.yaml to replace example values with your real values:

# Before (scaffold example values)
config:
  endpoints:
    - url: "https://example.com"
      expected_body: ""
    - url: "https://example.com/api/health"
      expected_body: ""


# After (your real values)
config:
  endpoints:
    - url: "https://acme.com"
      expected_body: ""
    - url: "https://acme.com/api/health"
      expected_body: ""
    - url: "https://acme.com/dashboard"
      expected_body: ""

Verify your configuration:

Terminal window
# Show current parameter values
tntc scaffold params show


# Verify no example values remain
tntc scaffold params validate

Alternative: Using —params-file

If you know all parameter values upfront, create a params.yaml and apply them during init:

params.yaml
endpoints:
  - url: "https://acme.com"
    expected_body: ""
  - url: "https://acme.com/api/health"
    expected_body: ""
latency_threshold_ms: 1500
probe_schedule: "*/2 * * * *"
Terminal window
tntc scaffold init uptime-tracker acme-uptime --params-file params.yaml

5. Set Up Secrets

Section titled “5. Set Up Secrets”
Terminal window
tntc secrets init

Edit .secrets.yaml with your actual secret references. Exoskeleton-managed services (like tentacular-postgres) have their credentials injected automatically at deploy time.

6. Validate and Test

Section titled “6. Validate and Test”
Terminal window
tntc validate
tntc test
tntc dev  # local dev server

7. Deploy

Section titled “7. Deploy”
Terminal window
tntc deploy

After deployment, verify:

Terminal window
tntc run acme-uptime
tntc status acme-uptime

What You Get

Section titled “What You Get”

After completing these steps, you have:

  • A tentacle at ~/tentacles/marketing-automations/acme-uptime/ with real configuration values
  • A tentacle.yaml recording that it came from uptime-tracker scaffold v1.0
  • A deployed and running workflow on your cluster in the marketing-automations enclave

Next Steps

Section titled “Next Steps”