truefoundry-helm（真实铸造头盔）

Routing note: For ambiguous user intents, use the shared clarification templates in references/intent-clarification.md.

Helm Chart Deployment

Deploy any Helm chart to TrueFoundry -- databases, caches, message queues, vector databases, monitoring tools, or any other OCI-compatible Helm chart. TrueFoundry supports any chart as long as the cluster can pull it from the registry.

Two paths:

1. CLI (tfy apply) -- Write a YAML manifest and apply it. Works everywhere.
2. REST API (fallback) -- When CLI unavailable, use tfy-api.sh.

When to Use

• User explicitly wants a Helm chart deployment for a database (PostgreSQL, MySQL, MongoDB, etc.)
• User wants to install a cache (Redis, Memcached)
• User wants to deploy a message queue (RabbitMQ, Kafka, NATS)
• User says "install helm chart", "deploy via helm"
• User wants infrastructure components, not application code
• User wants to deploy a vector database (Qdrant, Milvus, Weaviate, Chroma)
• User wants to deploy monitoring tools (Prometheus, Grafana)
• User has a custom/private Helm chart to deploy
• User wants to deploy ANY infrastructure component available as a Helm chart

If user intent is "deploy Postgres/Redis/database" without saying Helm, ask which strategy they want:

• Helm chart infrastructure (helm skill)
• Containerized service deployment (deploy skill)

When NOT to Use

• User wants to deploy application code -> prefer deploy skill; ask if the user wants another valid path
• User explicitly asks for Docker/container/image-based database deployment -> use deploy containerized service path (not Helm)
• User wants to check what's deployed -> prefer applications skill; ask if the user wants another valid path
• User wants to view logs -> prefer logs skill; ask if the user wants another valid path

Prerequisites

Always verify before deploying:

1. Credentials -- TFY_BASE_URL and TFY_API_KEY must be set (env or .env)
2. Workspace -- TFY_WORKSPACE_FQN required. Never auto-pick. Ask the user if missing.
3. CLI -- Check if tfy CLI is available: tfy --version. If not, pip install 'truefoundry==0.5.0'.

For credential check commands and.env setup, see references/prerequisites.md.

User Confirmation Checklist

Before deploying a Helm chart, ALWAYS confirm these with the user:

• Chart source -- Which chart? (suggest from common charts table)
• Chart registry -- Public (official chart registries) or private registry?
• Chart version -- Specific version or latest?
• Release name -- What to call this deployment? (default: chart name + random suffix)
• Namespace/Workspace -- Which workspace FQN? (never auto-pick)
• Environment -- Is this for dev, staging, or production? (affects resource defaults)
• Configuration -- Critical values to set:

- Passwords/credentials -- Use strong random values or reference TrueFoundry secrets - Storage size -- Persistent volume size (e.g., 10Gi, 20Gi) - Resources -- CPU/memory limits and requests - Replicas -- Number of instances (1 for dev, 3+ for prod) - Network -- Expose externally or internal-only?

• Auto-shutdown -- Should the deployment auto-stop after inactivity? (useful for dev/staging to save costs)

Do NOT deploy with minimal defaults without asking. Production databases need proper sizing, credentials, and persistence configuration.

Finding & Sourcing Helm Charts

For chart sources, OCI URLs, registries, version discovery, and the chart selection guide, see references/helm-chart-sources.md.

Key points: TrueFoundry supports oci-repo (recommended), helm-repo, and git-helm-repo source types. Do NOT use Bitnami charts. Always search Artifact Hub for the official chart from the project maintainers or use the chart publisher's own OCI registry.

Security: Helm charts from public registries are third-party code that runs in your cluster. Only use charts from trusted sources. Always pin chart versions — never use latest or floating tags. Review chart values before deploying. Verify the chart publisher is the official project maintainer. Do not allow the agent to discover and deploy charts from Artifact Hub without user confirmation of the chart source and version.

Deploy Flow

1. Gather Configuration

Ask the user for critical configuration values. For a PostgreSQL example:

I'll deploy PostgreSQL to TrueFoundry. Let me confirm a few things:

1. Chart version: Use postgresql 15.x (latest stable)? Or specific version?
2. Database name: What should the default database be called? (default: postgres)
3. Password: I'll generate a strong random password. Or do you have a TrueFoundry secret group to reference?
4. Storage: How much persistent storage? (default: 10Gi for dev, 50Gi+ for prod)
5. Resources:
   - CPU: 0.5 cores for dev, 2+ for prod?
   - Memory: 512Mi for dev, 2Gi+ for prod?
6. Replicas: 1 for dev, 3+ for prod high availability?
7. Access: Internal-only (default) or expose externally?

2. Generate YAML Manifest

Create a YAML manifest with user-confirmed values:

name: postgres-prod
type: helm
source:
  type: oci-repo
  version: "16.7.21"
  oci_chart_url: oci://REGISTRY/CHART_NAME  # Search Artifact Hub for the official chart
values:
  auth:
    postgresPassword: GENERATED_OR_SECRET_REF
    database: myapp
  primary:
    persistence:
      enabled: true
      size: 50Gi
    resources:
      requests:
        cpu: "2"
        memory: 2Gi
      limits:
        cpu: "4"
        memory: 4Gi
  readReplicas:
    replicaCount: 2
workspace_fqn: cluster-id:workspace-name

3. Write and Preview Manifest

Write the manifest to tfy-manifest.yaml:

tfy apply -f tfy-manifest.yaml --dry-run --show-diff

Show the preview output to the user.

4. Apply

After user confirms:

tfy apply -f tfy-manifest.yaml

Fallback: REST API

If tfy CLI is not available, convert the YAML manifest to JSON and deploy via REST API. See references/cli-fallback.md for the conversion process.

Important: The workspaceId must be the internal workspace ID (not the FQN). Get it from the workspaces skill: GET /api/svc/v1/workspaces?fqn=WORKSPACE_FQN -> use the id field.

When using direct API, set TFY_API_SH to the full path of this skill's scripts/tfy-api.sh. See references/tfy-api-setup.md for paths per agent.

Via Tool Call

tfy_applications_create_deployment(
    manifest={
        "name": "postgres-prod",
        "type": "helm",
        "source": {
            "type": "oci-repo",
            "version": "16.7.21",
            "oci_chart_url": "oci://REGISTRY/CHART_NAME"
        },
        "values": {...},
        "workspace_fqn": "cluster-id:workspace-name"
    },
    options={
        "workspace_id": "ws-internal-id",
        "force_deploy": false
    }
)

Note: This requires human approval (HITL) when using tool calls.

Via Direct API

TFY_API_SH=~/.claude/skills/truefoundry-helm/scripts/tfy-api.sh

# First, get workspace ID from FQN
$TFY_API_SH GET "/api/svc/v1/workspaces?fqn=${TFY_WORKSPACE_FQN}"

# Then deploy (JSON body)
$TFY_API_SH PUT /api/svc/v1/apps '{
  "manifest": {
    "name": "postgres-prod",
    "type": "helm",
    "source": {
      "type": "oci-repo",
      "version": "16.7.21",
      "oci_chart_url": "oci://REGISTRY/CHART_NAME"
    },
    "values": {
      "auth": {"postgresPassword": "...", "database": "myapp"},
      "primary": {
        "persistence": {"enabled": true, "size": "50Gi"},
        "resources": {
          "requests": {"cpu": "2", "memory": "2Gi"},
          "limits": {"cpu": "4", "memory": "4Gi"}
        }
      }
    },
    "workspace_fqn": "cluster-id:workspace-name"
  },
  "workspaceId": "WORKSPACE_ID_HERE"
}'

5. Report Connection Details

After successful deployment, provide the user with connection details (host, port, database, credentials). For connection DNS patterns and default ports by chart type, see references/helm-chart-sources.md (Connection Details by Chart section).

Example Configurations

For full YAML manifest examples (Redis, MongoDB, RabbitMQ, Qdrant, Elasticsearch), secrets management patterns, and environment-specific defaults, see references/helm-chart-examples.md.

Advanced: Kustomize & Additional Manifests

For Kustomize patches and deploying additional Kubernetes manifests alongside Helm charts, see references/helm-advanced.md.

After Deploy

After applying the Helm manifest, verify status automatically without asking an extra prompt.

Preferred verification path:

1. MCP tool call first:

tfy_applications_list(filters={"workspace_fqn": "WORKSPACE_FQN", "application_name": "RELEASE_NAME"})

1. Fallback to API:

$TFY_API_SH GET '/api/svc/v1/apps?workspaceFqn=WORKSPACE_FQN&applicationName=RELEASE_NAME'

Helm chart deployed successfully!

Next steps:
1. Deployment status verified and reported automatically
2. View logs: Use `logs` skill if there are issues
3. Connect from your app: Use the service DNS provided above
4. Store credentials: Use TrueFoundry secrets for app access

Addon Operations

TrueFoundry treats Helm deployments as "addons" and provides additional management endpoints.

Get Addon Details

# Get details about a deployed addon/Helm release
$TFY_API_SH GET /api/svc/v1/addon/ADDON_ID

Upgrade Addon

Upgrade an addon to the latest chart version:

# Upgrade to latest version
$TFY_API_SH POST /api/svc/v1/addon-upgrade/APPLICATION_ID

List Available Addon Components

See what addon types are available:

$TFY_API_SH GET /api/svc/v1/addon/list/components

<success_criteria>

Success Criteria

• The Helm chart is deployed and all pods are running in the target workspace
• The agent has confirmed the chart version, resource sizing, and credentials with the user before deploying
• Connection details (host, port, credentials) are provided to the user
• Deployment status is verified automatically immediately after apply (no extra prompt)
• Persistent storage is configured for stateful charts (databases, caches)
• The user can connect to the deployed service from their application using the provided DNS

</success_criteria>

Composability

• Find workspace first: Use workspaces skill to get workspace FQN and ID
• Check what's deployed: Use applications skill to list existing Helm releases
• Test after deployment: Use service-test skill to validate the deployed service
• Manage secrets: Use secrets skill to create secret groups before deploy
• View logs: Use logs skill with the HelmRelease application ID
• Connect from app: Reference the deployed chart's service DNS in your application's YAML manifest

Error Handling

For error messages and troubleshooting (workspace issues, chart not found, values validation, insufficient resources, PVC binding, connection issues), see references/helm-errors.md.

Additional CLI-specific errors:

• tfy: command not found -- Install with pip install 'truefoundry==0.5.0'
• tfy apply validation errors -- Check YAML syntax, ensure required fields (name, type, source, workspace_fqn) are present