PromptRenderResult
What a dry run rendered (1.55.0): the sections with their token estimates, the stable prefix, the names of missing values and the lint findings. Nothing is stored or audited.
An object with 7 fields.
Fields
| Field | Type | Required | Description |
|---|---|---|---|
lint | array of PromptLintFinding | yes | The lint findings of what was rendered: the stored findings of a version, those of the draft, and for an Agent version the cross-checks of its publish. |
missing_values | array of PromptValuePath | yes | The paths that rendered empty for want of a value, such as variables.clinician_name (missing_value); names only, never a value. |
object | "prompt_render" | yes | |
sections | array of PromptRenderSection | yes | In composition order. |
stable_prefix_tokens | integer | yes | The tokens before the first per-Call or per-Turn value, which a Provider that caches prompts may reuse between requests (prompt-manager.md section 4). |
total_tokens | integer | yes | |
variant | PromptVariantKey | yes |
Used by
The operations whose parameters, request body or responses name this shape directly.