Provider Ephemeral Resources logo

Provider Ephemeral Resources

Organization
hashicorp
provider-ephemeral-resources

Implement Terraform provider ephemeral resources with the Plugin Framework: the Open/Renew/Close lifecycle, ephemeral schema design, registration via EphemeralResources, renewal for expiring credentials, and how ephemeral values flow into write-only attributes and provider configuration. Use when adding an ephemeral resource, exposing secrets/tokens/certificates that must never persist in state or plan, deciding between an ephemeral resource and a data source, or wiring short-lived credentials from one provider into another.

Overview

Publisherhashicorp
Repositoryagent-skills
Skill nameprovider-ephemeral-resources
Stars
880
Forks
130
Bundled files
Instructions only
LicenseMPL-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by hashicorp on GitHub. Read the source before you install it.

Installation

Install the Provider Ephemeral Resources AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/hashicorp/agent-skills/tree/main/plugins/terraform/skills/provider-ephemeral-resources
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/hashicorp/agent-skills.git /tmp/agent-skills
mkdir -p .claude/skills
cp -r /tmp/agent-skills/plugins/terraform/skills/provider-ephemeral-resources .claude/skills/provider-ephemeral-resources
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Provider Ephemeral Resources in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Provider Ephemeral Resources on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Provider Ephemeral Resources is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Terraform Provider Ephemeral Resources

Ephemeral resources (Terraform 1.10+) produce values that are never persisted to state or plan. They exist for exactly one job: handing secrets — tokens, generated passwords, short-lived certificates, decrypted values — to the parts of a configuration that need them, without writing them to disk. Any data source that returns a sensitive value is a candidate to be (or to also exist as) an ephemeral resource.

Official docs: Ephemeral Resources.

When to Use One

SituationUse
Read-only lookup of non-sensitive dataData source
Value is sensitive and only needed at apply time (DB password for a provider block, token for a write-only attribute)Ephemeral resource
Sensitive value that downstream managed resources must store (e.g. as an attribute)Regular resource/data source — but pair with write-only attributes where possible
Credential that expires mid-operation (STS-style tokens, short-TTL leases)Ephemeral resource with Renew

Ephemeral results can be used in provider configuration, write-only attributes, provisioner configuration, and other ephemeral contexts — but not in regular attributes, because those persist to state.

Lifecycle

Terraform calls up to three methods per operation:

  • Open (required) — fetch or create the value; runs during plan and/or apply whenever the result is needed. There is no state to refresh and nothing to import.
  • Renew (optional) — called when the wall clock passes the RenewAt returned by Open/Renew, for values that expire while Terraform is still running. Renew cannot return a new result — it can only extend/refresh what Open produced (e.g. re-lease the same credential); if the value itself changes on renewal, the API is not renewable in this sense and Open must return a longer-lived value.
  • Close (optional) — called when Terraform is done with the value; revoke leases or delete temporary credentials here.

Open can pass bytes forward via resp.Private; Renew and Close receive them — use this for lease IDs needed to renew/revoke.

Implementation

go
var (
    _ ephemeral.EphemeralResource              = &tokenEphemeralResource{}
    _ ephemeral.EphemeralResourceWithConfigure = &tokenEphemeralResource{}
    _ ephemeral.EphemeralResourceWithRenew     = &tokenEphemeralResource{}
    _ ephemeral.EphemeralResourceWithClose     = &tokenEphemeralResource{}
)

func NewTokenEphemeralResource() ephemeral.EphemeralResource {
    return &tokenEphemeralResource{}
}

type tokenEphemeralResource struct {
    client *examplecloud.Client
}

type tokenEphemeralResourceModel struct {
    RoleName types.String `tfsdk:"role_name"`
    Token    types.String `tfsdk:"token"`
    LeaseID  types.String `tfsdk:"lease_id"`
}

func (r *tokenEphemeralResource) Metadata(_ context.Context, req ephemeral.MetadataRequest, resp *ephemeral.MetadataResponse) {
    resp.TypeName = req.ProviderTypeName + "_token"
}

func (r *tokenEphemeralResource) Schema(_ context.Context, _ ephemeral.SchemaRequest, resp *ephemeral.SchemaResponse) {
    resp.Schema = schema.Schema{
        Attributes: map[string]schema.Attribute{
            "role_name": schema.StringAttribute{
                Required:            true,
                MarkdownDescription: "Role to obtain a token for.",
            },
            "token": schema.StringAttribute{
                Computed:            true,
                Sensitive:           true,
                MarkdownDescription: "The issued token. Never persisted to state.",
            },
            "lease_id": schema.StringAttribute{
                Computed:            true,
                MarkdownDescription: "Identifier of the token lease.",
            },
        },
    }
}

func (r *tokenEphemeralResource) Open(ctx context.Context, req ephemeral.OpenRequest, resp *ephemeral.OpenResponse) {
    var data tokenEphemeralResourceModel
    resp.Diagnostics.Append(req.Config.Get(ctx, &data)...)
    if resp.Diagnostics.HasError() {
        return
    }

    lease, err := r.client.IssueToken(ctx, data.RoleName.ValueString())
    if err != nil {
        resp.Diagnostics.AddError(
            "Error opening Token",
            fmt.Sprintf("issuing token for role (%s): %s", data.RoleName.ValueString(), err),
        )
        return
    }

    data.Token = types.StringValue(lease.Token)
    data.LeaseID = types.StringValue(lease.ID)

    resp.RenewAt = lease.ExpiresAt.Add(-2 * time.Minute) // renew with margin
    resp.Private.SetKey(ctx, "lease_id", []byte(lease.ID))
    resp.Diagnostics.Append(resp.Result.Set(ctx, &data)...)
}

func (r *tokenEphemeralResource) Renew(ctx context.Context, req ephemeral.RenewRequest, resp *ephemeral.RenewResponse) {
    leaseID, diags := req.Private.GetKey(ctx, "lease_id")
    resp.Diagnostics.Append(diags...)
    if resp.Diagnostics.HasError() {
        return
    }

    lease, err := r.client.RenewLease(ctx, string(leaseID))
    if err != nil {
        resp.Diagnostics.AddError("Error renewing Token", err.Error())
        return
    }
    resp.RenewAt = lease.ExpiresAt.Add(-2 * time.Minute)
}

func (r *tokenEphemeralResource) Close(ctx context.Context, req ephemeral.CloseRequest, resp *ephemeral.CloseResponse) {
    leaseID, diags := req.Private.GetKey(ctx, "lease_id")
    resp.Diagnostics.Append(diags...)
    if resp.Diagnostics.HasError() {
        return
    }

    if err := r.client.RevokeLease(ctx, string(leaseID)); err != nil {
        resp.Diagnostics.AddError("Error closing Token", err.Error())
    }
}

Configure follows the same ProviderData-cast pattern as resources (the provider-resources skill, if available, shows it); the client comes from resp.EphemeralResourceData set in the provider's Configure.

Registration

The provider opts in via provider.ProviderWithEphemeralResources:

go
var _ provider.ProviderWithEphemeralResources = &examplecloudProvider{}

func (p *examplecloudProvider) EphemeralResources(_ context.Context) []func() ephemeral.EphemeralResource {
    return []func() ephemeral.EphemeralResource{
        NewTokenEphemeralResource(),
    }
}

Set resp.EphemeralResourceData = client in the provider's Configure alongside ResourceData/DataSourceData.

Design Rules

  • Never log the value, never put it in a diagnostic. The whole point is non-persistence; an error message containing the token defeats it.
  • Mark the secret attribute Sensitive: true anyway — it guards rendering in the ephemeral value's own lifecycle output.
  • No plan modifiers, no import, no id convention — there is no state for any of them to act on.
  • Schema inputs follow the same rules as data source arguments; expose the API's identifiers (role_name), not invented ones.
  • Set RenewAt with a safety margin before the real expiry; Terraform renews lazily, not on a precise timer.
  • If the upstream value cannot be revoked, skip Close rather than implementing a no-op that suggests revocation happens.

Testing

Ephemeral results never reach state, so tests assert them indirectly — the standard pattern echoes the ephemeral value through the echoprovider into a regular resource the test can inspect. Minimum coverage: a basic open-and-use test and per-attribute tests alongside required fields. Use the provider-test-patterns skill (if available) — its ephemeral testing reference covers the echoprovider setup, version gating (tfversion.SkipBelow(tfversion.Version1_10_0)), and multi-step patterns.

Documentation

Registry docs live at docs/ephemeral-resources/<name>.md, generated by tfplugindocs like every other page type. Use the provider-docs skill (if available) for the workflow; document the renewal/revocation behavior explicitly — users need to know whether closing their Terraform run revokes the credential.

Checklist

  • Value genuinely must not persist (otherwise a data source is simpler)
  • Open implemented; Renew/Close only where the API supports them
  • Secret attributes Sensitive: true; value never logged or in diagnostics
  • Lease/handle passed via Private, not via the result
  • RenewAt set with margin for expiring credentials
  • Registered in EphemeralResources(); EphemeralResourceData set in provider Configure
  • Echo-provider acceptance tests, version-gated to Terraform >= 1.10
  • Docs page explains lifetime, renewal, and revocation behavior

Frequently asked questions

What does the Provider Ephemeral Resources AI skill do?

Implement Terraform provider ephemeral resources with the Plugin Framework: the Open/Renew/Close lifecycle, ephemeral schema design, registration via EphemeralResources, renewal for expiring credentials, and how ephemeral values flow into write-only attributes and provider configuration. Use when adding an ephemeral resource, exposing secrets/tokens/certificates that must never persist in state or plan, deciding between an ephemeral resource and a data source, or wiring short-lived credentials from one provider into another.

Why use Provider Ephemeral Resources on TypingMind?

Because you install it once and use it with any model. Provider Ephemeral Resources is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Provider Ephemeral Resources in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/hashicorp/agent-skills/tree/main/plugins/terraform/skills/provider-ephemeral-resources. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Provider Ephemeral Resources?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Provider Ephemeral Resources?

As many as you like. As long as a model supports skills, you can use Provider Ephemeral Resources with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Provider Ephemeral Resources AI skill free?

Yes. It is published on GitHub by hashicorp under the MPL-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇