CLI commands and workflows
Install the draw.io prototype from its pinned GitHub commit until npm publication:
npm install github:architecture-tokens/drawio-prototype#f9e00c7c12750fe3c6fd576c48ca01b18ffcde62
npx archtokens validate payments.yaml --format human
npx archtokens generate payments.yaml --out payments.drawio --format humanThe CLI loads the pinned core, security, environment, and lifecycle libraries automatically and has no config files. Repeat --library for additional local token libraries and --policy for local policy sets. --format accepts human or json. For generate, --ai-model takes precedence over ARCHTOKENS_OPENAI_MODEL, then gpt-5.6; OPENAI_API_KEY is only used by generate. The offline golden workflow is the reproducible baseline, and no live API call is run here.
| Exit code | Meaning |
|---|---|
| 0 | Valid model and generated artifact |
| 1 | Invalid architecture |
| 2 | Usage or input error |
| 3 | API error |
| 4 | Invalid layout |
| 5 | Output error |
Live OpenAI verification is explicitly skipped in v0.1 and must never be required for offline builds.
Treat models as potentially sensitive: keep credentials out of YAML, prefer offline validation, and review generated artifacts before sharing.