メニュー
Routes AI agents through terraform-ingest workflows: install, configure, ingest, and module discovery. Use when the user mentions terraform-ingest, setting up a Terraform module library, ingesting modules, or getting started with this project.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Skill for the Terraform_ingest area of terraform-ingest. 261 symbols across 23 files.
Skill for the Tests area of terraform-ingest. 112 symbols across 15 files.
Builds and validates terraform-ingest config.yaml: repositories, output paths, embeddings, and MCP settings. Use when the user wants to configure terraform-ingest, add repositories, edit config.yaml, or enable vector search.
Runs and monitors terraform-ingest repository ingestion from YAML config, verifies JSON output, and interprets progress status. Use when the user wants to ingest modules, run first ingestion, refresh the module index, or check ingestion progress.
Mandatory before calling work complete when the session changed repo files. Use qa:quick during iteration and qa:prepush before hand-off. Read .terraform-ingest/summary.json only — not log files.
Installs terraform-ingest and prepares a fresh environment: dependencies, config initialization, and git credential checks. Use when the user wants to install terraform-ingest, get started, set up for the first time, or prepare before configuration and ingestion.
| name | terraform-ingest |
| description | Routes AI agents through terraform-ingest workflows: install, configure, ingest, and module discovery. Use when the user mentions terraform-ingest, setting up a Terraform module library, ingesting modules, or getting started with this project. |
Use this skill to pick the right workflow skill and interface (MCP, CLI, or API).
| User intent | Skill |
|---|---|
| Install, first-time setup, prerequisites | terraform-ingest-setup |
| Edit config.yaml, add repos, enable embeddings | terraform-ingest-configure |
| Run ingestion, monitor progress, verify output | terraform-ingest-ingest |
| Connect an agent to MCP (Phase 2) | See docs/mcp.md |
| Find or recommend modules (Phase 2) | Use MCP search_modules_vector after ingest |
1. terraform-ingest-setup → install + init config.yaml
2. terraform-ingest-configure → add repositories, validate
3. terraform-ingest-ingest → run ingest, verify output
4. (optional) docs/mcp.md → connect MCP for module queries
User request
│
├─ MCP connected? ──yes──► MCP tools (run_ingestion, get_ingestion_status, search_*)
│
└─ no ──► Shell access? ──yes──► terraform-ingest CLI
│
└─ no ──► Guide user with manual steps; API via curl if HTTP server running
| Task | MCP tool | CLI command |
|---|---|---|
| Start ingest | run_ingestion | terraform-ingest ingest config.yaml |
| Background ingest | run_ingestion(background=true) | terraform-ingest ingestion run --background |
| Poll progress | get_ingestion_status | terraform-ingest ingestion status |
| Add repository | — | terraform-ingest config add-repo --url ... |
| Validate config | — | terraform-ingest skills validate-config config.yaml |
| Deploy skills | — | terraform-ingest skills install --target cursor |
| Variable | Purpose | Default |
|---|---|---|
TERRAFORM_INGEST_CONFIG | Path to YAML config | config.yaml |
TERRAFORM_INGEST_OUTPUT_DIR | JSON summaries directory | ./output |
GITHUB_TOKEN | GitHub import / private repos | — |
OPENAI_API_KEY | OpenAI embedding strategy | — |
config commands over hand-editing YAML unless the user explicitly wants to edit the file.mcp.instructions from config — point users to customize that field after MCP is connected.