HashiCorp Agent Skills
A HashiCorp-maintained collection of 20 Agent Skills covering Terraform configuration, provider development and testing, policy, plus Packer workflows for AWS, Azure, Windows images, and HCP Registry.
Project overview
HashiCorp Agent Skills is a HashiCorp-maintained collection for Terraform and Packer. It breaks infrastructure style, provider development and testing, policy, and image-building tasks into 20 inspectable SKILL.md files while supporting both selective installation and product bundles. It is an Agent guidance and workflow layer, not Terraform, Packer, or a cloud control plane; real infrastructure, credentials, and spending still require explicit human approval and least-privilege controls.
Core capabilities
Terraform configuration and modules
terraform-style-guide, terraform-stacks, terraform-search-import, and refactor-module cover HCL generation and review, Stacks, existing-resource import, and modular refactoring.
Terraform provider development
Provider Skills cover new providers, configuration, resources, actions, ephemeral resources, Plugin SDK migration, and Registry documentation.
Testing and acceptance
terraform-test, provider-test-patterns, and run-acceptance-tests address Terraform tests, provider unit and integration patterns, and real acceptance tests.
Policy as code
terraform-policy writes, tests, or converts .policy.hcl and .policytest.hcl files, including Sentinel-to-tfpolicy migration.
Multi-cloud Packer images
Packer Skills guide AWS AMI, Azure managed and Compute Gallery images, and WinRM plus PowerShell-driven Windows builds.
HCP Packer Registry
The push-to-registry Skill guides publishing Packer build metadata to HCP Packer Registry for image lifecycle tracking and governance.
Installation and usage
Choose one task-specific Skill or a Terraform or Packer product bundle, then use only an installation path documented in the README. Verify the installed name and source with a credential-free, read-only, or static code task. Cloud credentials, real resources, acceptance tests, and image builds require separate approval.
Let an AI Agent install it
Send this prompt to Codex, Claude Code, or another AI agent that can work with your local environment.
Install HashiCorp Agent Skills from https://github.com/hashicorp/agent-skills. First read README.md, SKILLS.md, SECURITY.md, SUPPORT.md, and the target SKILL.md, then explain whether this host should use individual Skills CLI installation or the Terraform/Packer marketplace bundle. Install only the smallest Skill needed for the current task by default; if using Skills CLI, run only the npx skills add command documented by the repository. Report the actual write location, source, and installed name, then verify loading with a static, read-only, credential-free task. Before editing existing files, running Terraform/Packer/Go/tfpolicy, reading credentials, calling cloud APIs, creating real resources, running acceptance tests, building images, incurring cost, or risking leftover resources, explain the command, account, region, permissions, cost, and cleanup plan and ask me to approve. Do not treat the supported-model list as an outcome guarantee.- Selective installation needs a Node.js environment capable of running npx; no minimum Node.js version is documented
- Network access to github.com
- A host supporting Agent Skills or one of the documented plugin marketplaces
- Terraform, Packer, Go, or tfpolicy only when the selected task actually needs them
- Least-privilege AWS, Azure, HCP, or provider credentials only for authorized real-cloud tasks
npx skills add hashicorp/agent-skills- 1Choose the product and smallest Skill
Use SKILLS.md to find the Terraform or Packer entry matching the task and read its SKILL.md; do not install all 20 merely because the catalog is compact.
- 2Match the host installation path
General Agent Skills hosts use npx skills add, Claude Code can use HashiCorp marketplace commands, and Codex follows the .agents marketplace documented in the README.
- 3Run a repository-documented command
Start with npx skills add hashicorp/agent-skills to browse, then use the full README path for the reviewed Terraform or Packer Skill if selective installation is desired.
- 4Inspect names, source, and writes
Have the host list the result, confirm hashicorp/agent-skills as the source and the expected product and Skill names, and record configuration or filesystem writes.
- 5Perform a minimal static check
Ask the Agent to review sample HCL or explain the target Skill's checks; separately approve credentials, spending, and cleanup before real Terraform/Packer, acceptance-test, or cloud-build work.
After installation, ask the host to list installed Skills or plugins and verify the HashiCorp source, product bundle, and exact Skill name. Use Terraform formatting advice, static code review, or sample generation for the first check; do not run apply, acceptance tests, packer build, or any command that accesses a cloud account.
- npx skills add hashicorp/agent-skills lists and selects Skills; the README also documents complete paths for individual installs
- terraform@hashicorp and packer@hashicorp are Claude Code product bundles, not individual Skills
- The Codex marketplace path is in .agents/plugins/marketplace.json; the README does not publish a universal shell install command
- Product plugin manifests declare Read and Write, so least-privilege host permissions still matter after installation
- For pinned production workflows, record the reviewed commit SHA or assess v1.0.0 and subsequent main-branch changes
Use cases
Terraform HCL generation and review
Generate configurations against official style guidance, inspect variables and module boundaries, and structure Stacks, Search and Import, or refactoring work.
Terraform provider development
Plan and implement provider configuration, resources, actions, ephemeral resources, framework migration, and Registry documentation.
Testing and policy
Write Terraform tests, provider test patterns, and Policy files while reserving real acceptance tests for approved accounts, credentials, and cost controls.
Multi-cloud image builds
Prepare Packer configuration for AWS AMIs, Azure images, and Windows templates with explicit networking, authentication, timeout, and cleanup controls.
Image lifecycle governance
Connect Packer build metadata to HCP Packer Registry for version tracking and image lifecycle management.
Assessment
The collection's main value comes from HashiCorp publisher identity, 20 individually inspectable Skills, both selective and product-bundle installation, and clear security, support, and model-scope documentation. It concentrates on high-value Terraform and Packer engineering workflows and has maintenance commits after v1.0.0. Its boundaries are equally important: Skills are guidance, product manifests have Read and Write capabilities, and acceptance tests or image builds touch real credentials, resources, and spending. This review verified repository evidence and manifests but did not install or test functions in a cloud environment.
Why it may be useful
- Clear HashiCorp publisher identity and MPL-2.0 license
- Twenty real SKILL.md files match the catalog count
- Focused, inspectable Terraform and Packer workflow coverage
- Skills CLI plus Claude Code and Codex marketplace paths
- Complete security, support, model-scope, changelog, and release records
What to know first
- Collection review does not replace per-Skill command and permission review
- Supported model scope does not ensure identical outcomes across hosts
- Real acceptance tests and image builds can incur costs or leave resources
- Installation, functions, compatibility, and cloud execution were not tested
README
HashiCorp Agent Skills
Overview
A HashiCorp-maintained collection of 20 Agent Skills covering Terraform configuration, provider development and testing, policy, plus Packer workflows for AWS, Azure, Windows images, and HCP Registry. A collection of Agent skills and Claude Code plugins for HashiCorp products.
Getting started
- Choose one task-specific Skill or a Terraform or Packer product bundle, then use only an installation path documented in the README. Verify the installed name and source with a credential-free, read-only, or static code task. Cloud credentials, real resources, acceptance tests, and image builds require separate approval.
- The README presents the repository as HashiCorp Agent Skills for Terraform and Packer and lists 20 Skills.
- The 16 Terraform Skills cover HCL, Stacks, testing, Search and Import, Policy, modules, and Plugin Framework development.
- After installation, ask the host to list installed Skills or plugins and verify the HashiCorp source, product bundle, and exact Skill name. Use Terraform formatting advice, static code review, or sample generation for the first check; do not run apply, acceptance tests, packer build, or any command that accesses a cloud account.
Configuration
npx skills add hashicorp/agent-skills