Eval Types

`CallerPermissions`

Evaluated permissions for the current caller on this artifact type.

Field	Type	Required	Description
`can_create`	`boolean`	Yes	Whether the caller can create new artifacts
`can_draft`	`boolean`	Yes	Whether the caller can create/update drafts
`can_duplicate`	`boolean`	Yes	Whether the caller can duplicate artifacts
`has_access`	`boolean`	No	Whether the caller can view this entity
`can_edit`	`boolean`	No	Whether the caller can edit this entity
`can_delete`	`boolean`	No	Whether the caller can delete this entity
`disabled_reason`	`string`	No	Human-readable reason if editing is disabled

`ColumnInfo`

Field	Type	Required	Description
`name`	`string`	Yes	Column name
`type`	`string`	Yes	Column data type
`nullable`	`boolean`	Yes	Whether the column is nullable

`CreateEvalItem`

Single eval item for create — no eval_id.

Required fields (name): provide ID or value.

Field	Type	Required	Description
`id`	`string`	No	Optional pre-assigned UUID
`resource_id`	`string`	No	Optional preset UUID for the resource snapshot
`name_id`	`string`	No	REQUIRED FOR CREATE (or pass `name`): UUID of an existing name resource
`name`	`string`	No	REQUIRED FOR CREATE (or pass `name_id`): display name text — creates a new name resource if `name_id` is not provided
`description_id`	`string`	No	Description resource UUID
`description`	`string`	No	Description value for resolution
`flag_ids`	`string`[]	No	Flag option UUIDs
`department_ids`	`string`[]	No	Department UUIDs
`departments`	`string`[]	No	Department names for resolution
`model_ids`	`string`[]	No	Model UUIDs
`model_flag_ids`	`string`[]	No	Model flag UUIDs
`model_rubric_ids`	`string`[]	No	Model rubric UUIDs
`model_position_ids`	`string`[]	No	Model position UUIDs
`active`	`boolean`	No	Denormalized eval_active flag state; resolved to a flag_ids entry server-side

`DeleteEvalResult`

Per-item result within a bulk delete response.

Field	Type	Required	Description
`success`	`boolean`	Yes	Whether the operation succeeded
`eval_id`	`string`	No	Eval UUID (None only for soft-skipped rows)
`message`	`string`	Yes	Human-readable result message

`DocsApiResponse`

Field	Type	Required	Description
`list`	`PageMetaItem`	Yes	—
`detail`	`PageMetaItem`	Yes	—
`new`	`PageMetaItem`	Yes	—

`DocsResponse`

Field	Type	Required	Description
`name`	`string`	Yes	Resource or entry name
`type`	`string`	Yes	Resource or entry type identifier
`description`	`string`	Yes	Human-readable description
`materialized_view`	`MvInfo`	No	Materialized view metadata
`tables`	`TableInfo`[]	Yes	Related database tables
`operations`	`OperationInfo`[]	Yes	Available operations

`EvalDepartmentResource`

Field	Type	Required	Description
`department_id`	`string`	No	Department identifier
`name`	`string`	No	Display name
`description`	`string`	No	Description text
`generated`	`boolean`	No	Whether this was AI-generated
`suggested`	`boolean`	No	Whether this is a suggested option
`selected`	`boolean`	No	Whether this is currently selected
`pending`	`boolean`	No	Whether this selection is pending acceptance

`EvalDescriptionResource`

Field	Type	Required	Description
`id`	`string`	No	Unique identifier
`description`	`string`	No	Description text
`generated`	`boolean`	No	Whether this was AI-generated
`suggested`	`boolean`	No	Whether this is a suggested option
`selected`	`boolean`	No	Whether this is currently selected
`pending`	`boolean`	No	Whether this selection is pending acceptance

`EvalFieldError`

Per-field error from value resolution.

Field	Type	Required	Description
`field`	`string`	Yes	Field name that has the error
`message`	`string`	Yes	Human-readable error message

`EvalFlagResource`

Flag option row — one per (name, type, value) entry in flags_resource.

Field	Type	Required	Description
`id`	`string`	No	Flag resource identifier
`name`	`string`	No	Flag display name
`type`	`string`	No	Flag type (e.g. ‘eval_active’)
`value`	`boolean`	No	Underlying bool value of this option
`description`	`string`	No	Flag description text
`icon_id`	`string`	No	Icon identifier for the flag
`icon`	`string`	No	Resolved SVG markup (hydrated from icons_resource)
`generated`	`boolean`	No	Whether the flag was AI-generated
`suggested`	`boolean`	No	Whether this item is suggested
`selected`	`boolean`	No	Whether this item is selected
`pending`	`boolean`	No	Whether this item is pending acceptance

`EvalModelFlagOptionResource`

Cross-product option row: one per (model_id, flag_type, value) tuple. The ModelFlags picker groups these by (model_id, type) and renders a Switch per group; toggling picks the row whose value matches the new state and sends its flag_id (or denormalized {model_id,type,value}) up through the draft endpoint.

Field	Type	Required	Description
`model_id`	`string`	No	Model identifier
`flag_id`	`string`	No	Flag resource identifier (flags_resource row)
`type`	`string`	No	Flag type, e.g. ‘model_active’
`value`	`boolean`	No	Underlying flag value for this option
`name`	`string`	No	Display name from the flags_resource row
`description`	`string`	No	Description text from the flags_resource row
`icon`	`string`	No	Icon SVG markup hydrated from icons_resource

`EvalModelFlagResource`

Field	Type	Required	Description
`id`	`string`	No	Model-flag resource identifier
`model_id`	`string`	No	Associated model identifier
`flag_id`	`string`	No	Associated flag identifier
`type`	`string`	No	Flag type (e.g. ‘model_active’) of the linked flags_resource row
`value`	`boolean`	No	Underlying bool value of the linked flags_resource row
`name`	`string`	No	Display name
`description`	`string`	No	Description text
`icon`	`string`	No	Icon identifier
`generated`	`boolean`	No	Whether this was AI-generated
`suggested`	`boolean`	No	Whether this is a suggested option
`selected`	`boolean`	No	Whether this is currently selected
`pending`	`boolean`	No	Whether this selection is pending acceptance

`EvalModelFlagValue`

Denormalized per-(model, type) selection. Clients send these to update model-flag selections without needing the underlying flag_id; the server resolves (type, value) -> flag_id and upserts the model_flags_resource junction row.

Field	Type	Required	Description
`model_id`	`string`	Yes	Target model identifier
`type`	`string`	Yes	Flag type, e.g. ‘model_active’
`value`	`boolean`	Yes	Desired flag value

`EvalModelPositionResource`

Field	Type	Required	Description
`id`	`string`	No	Model-position resource identifier
`model_id`	`string`	No	Associated model identifier
`value`	`integer` \| `number`	No	Associated position value
`generated`	`boolean`	No	Whether this was AI-generated
`suggested`	`boolean`	No	Whether this is a suggested option
`selected`	`boolean`	No	Whether this is currently selected
`pending`	`boolean`	No	Whether this selection is pending acceptance

`EvalModelResource`

Field	Type	Required	Description
`id`	`string`	No	Model resource identifier
`name`	`string`	No	Display name
`description`	`string`	No	Description text
`modality_ids`	`string`[]	No	Associated modality identifiers
`generated`	`boolean`	No	Whether this was AI-generated
`suggested`	`boolean`	No	Whether this is a suggested option
`selected`	`boolean`	No	Whether this is currently selected
`pending`	`boolean`	No	Whether this selection is pending acceptance

`EvalModelRubricResource`

Field	Type	Required	Description
`id`	`string`	No	Model-rubric resource identifier
`model_id`	`string`	No	Associated model identifier
`rubric_id`	`string`	No	Associated rubric identifier
`generated`	`boolean`	No	Whether this was AI-generated
`suggested`	`boolean`	No	Whether this is a suggested option
`selected`	`boolean`	No	Whether this is currently selected
`pending`	`boolean`	No	Whether this selection is pending acceptance

`EvalNameResource`

Field	Type	Required	Description
`id`	`string`	No	Unique identifier
`name`	`string`	No	Display name
`generated`	`boolean`	No	Whether this was AI-generated
`suggested`	`boolean`	No	Whether this is a suggested option
`selected`	`boolean`	No	Whether this is currently selected
`pending`	`boolean`	No	Whether this selection is pending acceptance

`EvalResultItem`

Per-item result within a bulk create/update response.

Field	Type	Required	Description
`success`	`boolean`	Yes	Whether the operation succeeded
`eval_id`	`string`	No	Eval UUID
`message`	`string`	Yes	Human-readable result message
`errors`	`EvalFieldError`[]	No	List of per-field errors

`EvalRubricResource`

Top-level rubric catalog entry. The ModelRubrics picker renders one of these per option (alongside a “no rubric” clear state).

Field	Type	Required	Description
`id`	`string`	No	Rubric resource UUID
`name`	`string`	No	Rubric display name
`description`	`string`	No	Rubric description

`EvalSetup`

Run-level eval scaffold — first-class on the generate response.

Audit’s **output spread carries this onto <artifact>.generate.completed. Null when no rubric-bearing agent participated.

Field	Type	Required	Description
`test_id`	`string`	Yes	—
`invocations`	`InvocationSlot`[]	Yes	—

`GenerateConfig`

Developer configuration — all optional with sensible defaults.

Field	Type	Required	Description
`operations`	`string`[]	No	—
`dangerous`	`boolean`	No	—
`params`	`object`	No	—
`group_id`	`string`	No	—

`GenerationsEvalListItem`

Single generation group in the eval generations response.

Field	Type	Required	Description
`group_id`	`string`	Yes	UUID of the generation group
`session_id`	`string`	No	UUID of the parent session
`group_name`	`string`	No	Name of the generation group
`created_at`	`string`	No	Timestamp of the generation

`GetEvalDraftResponse`

Field	Type	Required	Description
`id`	`string`	Yes	UUID of the draft
`created_at`	`string`	Yes	Creation timestamp
`generated`	`boolean`	Yes	Whether this was AI-generated
`mcp`	`boolean`	Yes	Whether MCP tooling was used
`active`	`boolean`	Yes	Whether this draft is active
`session_id`	`string`	Yes	Associated session UUID
`name`	`string`	No	Immutable draft label set at create time
`department_ids`	`string`[]	Yes	Associated department UUIDs
`description_ids`	`string`[]	Yes	Associated description UUIDs
`flag_ids`	`string`[]	Yes	Associated flag UUIDs
`model_ids`	`string`[]	Yes	Associated model UUIDs
`name_ids`	`string`[]	Yes	Associated name UUIDs
`profile_ids`	`string`[]	Yes	Associated profile UUIDs
`rubric_ids`	`string`[]	Yes	Associated rubric UUIDs
`model_flag_ids`	`string`[]	No	Associated model flag UUIDs
`model_position_ids`	`string`[]	No	Associated model position UUIDs
`model_rubric_ids`	`string`[]	No	Associated model rubric UUIDs
`pending_department_ids`	`string`[]	No	Associated department UUIDs stored as inactive pending links
`pending_description_ids`	`string`[]	No	Associated description UUIDs stored as inactive pending links
`pending_flag_ids`	`string`[]	No	Associated flag UUIDs stored as inactive pending links
`pending_model_ids`	`string`[]	No	Associated model UUIDs stored as inactive pending links
`pending_name_ids`	`string`[]	No	Associated name UUIDs stored as inactive pending links
`pending_rubric_ids`	`string`[]	No	Associated rubric UUIDs stored as inactive pending links
`pending_model_flag_ids`	`string`[]	No	Pending model flag UUIDs
`pending_model_position_ids`	`string`[]	No	Pending model position UUIDs
`pending_model_rubric_ids`	`string`[]	No	Pending model rubric UUIDs

`GroupCall`

Tool call referenced by a message.

Field	Type	Required	Description
`id`	`string`	Yes	—
`tool_name`	`string`	No	—
`template_name`	`string`	No	—
`tool`	`object`	No	—
`ledger_status`	`string`	No	—
`ledger_operation`	`string`	No	—
`ledger_artifact`	`string`	No	—
`ledger_artifact_id`	`string`	No	—

`GroupMessage`

Message within a run.

Field	Type	Required	Description
`id`	`string`	Yes	—
`role`	`string`	Yes	—
`created_at`	`string`	No	—
`text_ids`	`string`[]	No	—
`audio_ids`	`string`[]	No	—
`image_ids`	`string`[]	No	—
`video_ids`	`string`[]	No	—
`file_ids`	`string`[]	No	—
`call_ids`	`string`[]	No	—
`calls`	`GroupCall`[]	No	—
`reasoning`	`boolean`	No	True when this row is a chain-of-thought trace persisted alongside the assistant answer (rendered as a collapsed accordion).
`in_context`	`boolean`	No	Whether this message is included in the LLM context for the next generation. Mirrors the dedup pass that builds chat history (see in_context_reason).
`in_context_reason`	`string`	No	Why this message is in/out of LLM context. ‘kept’ = included; ‘deduped_read’ = older read-only call to a tool that has a fresher result later in the group; future values may include ‘trimmed_top_n’.

`GroupResource`

Lightweight \{id, name\} for cross-referencing run-level ids (model_id / agent_id / profile_id) against human-readable names on the analytics panel. Names come from the canonical get_models / get_agents / get_profiles black boxes.

Field	Type	Required	Description
`id`	`string`	Yes	—
`name`	`string`	No	—

`GroupRun`

Run within a group, with its messages.

Carries token / cost / model / agent / profile attribution so the analytics view can render per-run cost + actor info without a parallel detail shape. profile_id is the authoring profile (human user), agent_id is the LLM-side actor, model_id is the model used by that agent. All optional — runs predating these columns or with unresolved attributions surface None.

Field	Type	Required	Description
`id`	`string`	Yes	—
`created_at`	`string`	No	—
`input_tokens`	`integer`	No	—
`output_tokens`	`integer`	No	—
`cached_input_tokens`	`integer`	No	—
`cost`	`number`	No	—
`model_id`	`string`	No	—
`agent_id`	`string`	No	—
`profile_id`	`string`	No	—
`previous_context_start_index`	`integer`	No	Index in `messages` where the current run’s own messages begin; earlier rows are previous-context replay. `None` when the run has no previous context attached.
`messages`	`GroupMessage`[]	No	—

`ImportField`

Field descriptor for CSV import column mapping.

Field	Type	Required	Description
`key`	`string`	Yes	—
`label`	`string`	Yes	—
`required`	`boolean`	No	—
`multi`	`boolean`	No	—
`type`	`string`	No	—
`example`	`string`	No	—
`description`	`string`	No	—

`InvocationSlot`

One agent’s slot in a multi-agent generation pool.

Populated by setup_generation_test when an agent carries a rubric. The client uses these IDs to drive the eval workflow: review the candidate’s output, optionally fire a grader against its invocation_id, and promote/reject by call_id via the existing idempotency_key + accept pattern.

Field	Type	Required	Description
`invocation_id`	`string`	Yes	—
`agent_id`	`string`	Yes	—
`rubric_id`	`string`	No	—

`ListEvalApiEval`

Eval type for list endpoint with computed permissions.

Field	Type	Required	Description
`id`	`string`	No	Eval artifact UUID (canonical id; mirrors eval_id)
`eval_id`	`string`	No	Eval UUID
`name`	`string`	No	Eval name
`description`	`string`	No	Eval description
`department_ids`	`string`[]	No	Associated department IDs
`model_ids`	`string`[]	No	Model resource UUIDs assigned to this eval
`rubric_ids`	`string`[]	No	Rubric artifact UUIDs assigned to this eval (via model_rubrics_resource)
`is_inactive`	`boolean`	No	Whether the eval is inactive
`is_dynamic`	`boolean`	No	Whether the eval uses dynamic mode
`use_groups`	`boolean`	No	Whether the eval uses groups
`pending_status`	`string`	No	Latest soft_calls_mv status: ‘pending’ / ‘accepted’ / ‘rejected’
`pending_operation`	`string`	No	Operation type (‘create'
`pending_call_id`	`string`	No	call_id (idempotency key for ack) of the pending op
`num_runs`	`integer`	No	Number of eval runs
`num_groups`	`integer`	No	Number of eval groups
`can_edit`	`boolean`	No	Whether the current user can edit
`can_duplicate`	`boolean`	No	Whether the current user can duplicate
`can_delete`	`boolean`	No	Whether the current user can delete
`updated_at`	`string`	No	Last updated timestamp

`ListFilterOption`

Standardized option for list endpoint filter sections.

Field	Type	Required	Description
`id`	`string`	No	Unique identifier for this filter option
`name`	`string`	No	Display name
`count`	`integer`	No	Number of matching records
`hex_code`	`string`	No	Hex color code for display
`value`	`string`	No	Internal value
`type`	`string`	No	Option type discriminator

`ListFilterSection`

Filter section with options and echoed request state.

Field	Type	Required	Description
`options`	`ListFilterOption`[]	No	Available filter options
`selected_ids`	`string`[]	No	Currently selected filter option IDs
`search`	`string`	No	Active search text for filtering

`MvInfo`

Field	Type	Required	Description
`name`	`string`	Yes	Materialized view name
`definition`	`string`	Yes	SQL definition of the view
`columns`	`ColumnInfo`[]	Yes	List of columns in the view

`OperationInfo`

Field	Type	Required	Description
`name`	`string`	Yes	Operation name
`description`	`string`	Yes	Human-readable description of the operation
`params`	`ParamInfo`[]	Yes	List of operation parameters
`returns`	`object`	No	Return type schema

`OperationPrompts`

Starter prompts keyed by operation name.

Each key is an operation (e.g. “create”, “search”, “draft”, “export”) and the value is a list of starter prompts for that operation. The client picks from the operations the caller has permission for and rotates through them.

Field	Type	Required	Description
`prompts`	`object`	No	Map of operation name to starter prompts

`PageMetaItem`

Field	Type	Required	Description
`title`	`string`	Yes	—
`description`	`string`	Yes	—

`ParamInfo`

Field	Type	Required	Description
`name`	`string`	Yes	Parameter name
`type`	`string`	Yes	Parameter data type
`required`	`boolean`	Yes	Whether the parameter is required
`default`	`any`	No	Default value if not required

`ProducedMedia`

One asset produced by a generation run.

resource_id is the canonical id the per-artifact download tools accept (e.g. Scenario_Image_Download(image_id=resource_id) for modality="image"). It maps to images_resource.id / videos_resource.id / audios_resource.id depending on the modality.

Field	Type	Required	Description
`modality`	`"image"` \| `"video"` \| `"audio"`	Yes	—
`resource_id`	`string`	Yes	—
`upload_id`	`string`	Yes	—
`mime_type`	`string`	No	—
`file_size`	`integer`	No	—

`ProfileSummary`

Caller identity derived from JWT — who you are on this page.

Superset of the old six-field version: now carries everything the client needs so that /\{artifact\}/context fully replaces /profiles/context and the extra getLayoutContextData round-trip can be dropped.

Field	Type	Required	Description
`name`	`string`	Yes	Display name of the authenticated user
`role`	`string`	Yes	Role name (e.g. ‘Super Administrator’)
`role_level`	`integer`	Yes	Role hierarchy level (0 = highest privilege)
`department_ids`	`string`[]	Yes	Departments the user belongs to
`artifact_access`	`string`[]	Yes	Artifact types this role can access (sidebar visibility)
`role_permissions`	`any`[][]	Yes	Full (artifact, operation) permission tuples for granular page gating
`is_active`	`boolean`	Yes	Whether the user’s profile is active
`id`	`string`	Yes	Profile UUID (SocketProvider, ProfileProvider)
`theme`	`ThemeBundle`	No	Resolved theme: hex primitives + derived oklch tokens + score thresholds
`session_id`	`string`	No	Current session UUID
`is_emulation`	`boolean`	No	Whether user is in emulation mode (ProfileProvider)
`role_resources`	`QGetProfileContextV4RoleResource`[]	No	All role resources for emulation display (ProfileProvider)
`scoped_roles`	`string`[]	No	Roles the user can emulate (ProfileProvider)
`active`	`boolean`	No	Alias for is_active (ProfileProvider uses this name)

`QGetProfileContextV4RoleResource`

Field	Type	Required	Description
`role`	`string`	No	—
`name`	`string`	No	—
`description`	`string`	No	—
`icon_value`	`string`	No	—
`color_hex`	`string`	No	—

`RunStatus`

Outcome of one watched run.

Field	Type	Required	Description
`run_id`	`string`	Yes	—
`status`	`"pending"` \| `"completed"` \| `"failed"`	Yes	—
`modality`	`string`	No	Modality of generated output (image, video, text…)
`resource_ids`	`string`[]	No	Resource UUIDs produced by this run (e.g. images_resource ids)
`error`	`string`	No	—

`TableInfo`

Field	Type	Required	Description
`name`	`string`	Yes	Table name
`columns`	`ColumnInfo`[]	Yes	List of columns in the table

`ThemeBundle`

Full theme payload for a page bootstrap.

Riding along on every /\{artifact\}/context response via ProfileSummary.theme. Layers:

primitives / dark_primitives — hex inputs the settings editor reads/writes (light + dark palettes).
tokens / dark_tokens — oklch tokens the client paints with. ThemeStyle emits two <style> blocks: one scoped to :root:not(.dark) (light) and one to :root.dark (dark).
thresholds — numeric score thresholds for analytics components. Empty-in → empty-out per token: missing values fall through to the matching globals.css default.

Field	Type	Required	Description
`primitives`	`ThemePrimitives`	No	Hex inputs from the setting (light palette, for the theme editor)
`tokens`	`ThemeTokens`	No	Derived oklch tokens for light mode (SSR CSS-var injection)
`dark_primitives`	`ThemePrimitives`	No	Hex inputs from the setting (dark palette, for the theme editor)
`dark_tokens`	`ThemeTokens`	No	Derived oklch tokens for dark mode (SSR CSS-var injection)
`thresholds`	`Thresholds`	No	Score thresholds resolved from the setting

`ThemePrimitives`

40 optional fields. The 17 essentials drive the rest; the other 23 are overrides for fine-tuning when derivation isn’t what you want.

Empty primitive → empty token → client falls back to globals.css.

Field	Type	Required	Description
`background`	`string`	No	—
`primary`	`string`	No	—
`accent`	`string`	No	—
`card`	`string`	No	—
`sidebar`	`string`	No	—
`muted_foreground`	`string`	No	—
`ring`	`string`	No	—
`border`	`string`	No	—
`destructive`	`string`	No	—
`success`	`string`	No	—
`warning`	`string`	No	—
`info`	`string`	No	—
`chart1`	`string`	No	—
`chart2`	`string`	No	—
`chart3`	`string`	No	—
`chart4`	`string`	No	—
`chart5`	`string`	No	—
`foreground`	`string`	No	—
`card_foreground`	`string`	No	—
`popover`	`string`	No	—
`popover_foreground`	`string`	No	—
`primary_foreground`	`string`	No	—
`secondary`	`string`	No	—
`secondary_foreground`	`string`	No	—
`muted`	`string`	No	—
`accent_foreground`	`string`	No	—
`destructive_foreground`	`string`	No	—
`danger`	`string`	No	—
`danger_foreground`	`string`	No	—
`input`	`string`	No	—
`success_foreground`	`string`	No	—
`warning_foreground`	`string`	No	—
`info_foreground`	`string`	No	—
`sidebar_foreground`	`string`	No	—
`sidebar_primary`	`string`	No	—
`sidebar_primary_foreground`	`string`	No	—
`sidebar_accent`	`string`	No	—
`sidebar_accent_foreground`	`string`	No	—
`sidebar_border`	`string`	No	—
`sidebar_ring`	`string`	No	—

`ThemeTokens`

40 fully-resolved CSS variable values (snake_case 1:1 with vars).

Field	Type	Required	Description
`background`	`string`	No	—
`foreground`	`string`	No	—
`card`	`string`	No	—
`card_foreground`	`string`	No	—
`popover`	`string`	No	—
`popover_foreground`	`string`	No	—
`primary`	`string`	No	—
`primary_foreground`	`string`	No	—
`secondary`	`string`	No	—
`secondary_foreground`	`string`	No	—
`muted`	`string`	No	—
`muted_foreground`	`string`	No	—
`accent`	`string`	No	—
`accent_foreground`	`string`	No	—
`destructive`	`string`	No	—
`destructive_foreground`	`string`	No	—
`danger`	`string`	No	—
`danger_foreground`	`string`	No	—
`border`	`string`	No	—
`input`	`string`	No	—
`ring`	`string`	No	—
`success`	`string`	No	—
`success_foreground`	`string`	No	—
`warning`	`string`	No	—
`warning_foreground`	`string`	No	—
`info`	`string`	No	—
`info_foreground`	`string`	No	—
`chart1`	`string`	No	—
`chart2`	`string`	No	—
`chart3`	`string`	No	—
`chart4`	`string`	No	—
`chart5`	`string`	No	—
`sidebar`	`string`	No	—
`sidebar_foreground`	`string`	No	—
`sidebar_primary`	`string`	No	—
`sidebar_primary_foreground`	`string`	No	—
`sidebar_accent`	`string`	No	—
`sidebar_accent_foreground`	`string`	No	—
`sidebar_border`	`string`	No	—
`sidebar_ring`	`string`	No	—

`Thresholds`

Numeric score thresholds resolved from the active setting.

Server pre-buckets dashboard metrics into success | warning | danger | neutral already, so most components don’t need these values. Surface them for chart reference lines, tooltips, and any client-side bucketing.

Field	Type	Required	Description
`success`	`integer`	Yes	Score >= this counts as success
`warning`	`integer`	Yes	Score >= this counts as warning
`danger`	`integer`	Yes	Score < success threshold but >= this counts as danger; below is neutral/no-data

`UpdateEvalItem`

Single eval item for update — eval_id required, all fields optional.

Only provided fields are updated (partial update).

Field	Type	Required	Description
`id`	`string`	Yes	Eval UUID to update
`name_id`	`string`	No	Name resource UUID
`name`	`string`	No	Name value for resolution
`description_id`	`string`	No	Description resource UUID
`description`	`string`	No	Description value for resolution
`flag_ids`	`string`[]	No	Flag option UUIDs
`department_ids`	`string`[]	No	Department UUIDs
`departments`	`string`[]	No	Department names for resolution
`model_ids`	`string`[]	No	Model UUIDs
`model_flag_ids`	`string`[]	No	Model flag UUIDs
`model_rubric_ids`	`string`[]	No	Model rubric UUIDs
`model_position_ids`	`string`[]	No	Model position UUIDs
`active`	`boolean`	No	Denormalized eval_active flag state; resolved to a flag_ids entry server-side

`UpdateEvalPatch`

Shared patch for bulk-update-all-matching mode.

Inherits every field from UpdateEvalItem and just relaxes id to optional — the bulk impl stamps the resolved id onto a clone of the patch per matched row, so any client-supplied id is ignored. Sparse semantics: only fields the client sets are written.

Field	Type	Required	Description
`id`	`string`	No	Ignored — bulk impl stamps the resolved eval id per matched row
`name_id`	`string`	No	Name resource UUID
`name`	`string`	No	Name value for resolution
`description_id`	`string`	No	Description resource UUID
`description`	`string`	No	Description value for resolution
`flag_ids`	`string`[]	No	Flag option UUIDs
`department_ids`	`string`[]	No	Department UUIDs
`departments`	`string`[]	No	Department names for resolution
`model_ids`	`string`[]	No	Model UUIDs
`model_flag_ids`	`string`[]	No	Model flag UUIDs
`model_rubric_ids`	`string`[]	No	Model rubric UUIDs
`model_position_ids`	`string`[]	No	Model position UUIDs
`active`	`boolean`	No	Denormalized eval_active flag state; resolved to a flag_ids entry server-side

`appinfraevaltypesDraftFormState`

Server-authoritative form state returned after draft save.

Field	Type	Required	Description
`name_id`	`string`	No	Selected name resource UUID
`name`	`string`	No	Echoed selected name value
`description_id`	`string`	No	Selected description resource UUID
`description`	`string`	No	Echoed selected description value
`flag_ids`	`string`[]	No	Selected flag option UUIDs
`active`	`boolean`	No	Echoed eval_active flag state
`department_ids`	`string`[]	No	Selected department UUIDs
`model_ids`	`string`[]	No	Selected model UUIDs
`model_flag_ids`	`string`[]	No	Selected model flag UUIDs
`model_flag_values`	`EvalModelFlagValue`[]	No	Denormalized (model_id, type, value) echo derived from model_flag_ids
`model_position_ids`	`string`[]	No	Selected model position UUIDs
`model_rubric_ids`	`string`[]	No	Selected model rubric UUIDs
`pending_ids`	`string`[]	No	Pending resource identifiers

`appinfraevaltypesSectionFilter`

Field	Type	Required	Description
`search`	`string`	No	Filter options by search text
`limit`	`integer`	No	Max options to return
`selected`	`boolean`	No	Only return selected items
`suggested`	`boolean`	No	Only return suggested items
`include`	`boolean`	No	Include this section in response (default true)
`parameter_ids`	`string`[]	No	Reserved for compatibility with shared filter parsing

Eval Types

CallerPermissions

ColumnInfo

CreateEvalItem

DeleteEvalResult

DocsApiResponse

DocsResponse

EvalDepartmentResource

EvalDescriptionResource

EvalFieldError

EvalFlagResource

EvalModelFlagOptionResource

EvalModelFlagResource

EvalModelFlagValue

EvalModelPositionResource

EvalModelResource

EvalModelRubricResource

EvalNameResource

EvalResultItem

EvalRubricResource

EvalSetup

GenerateConfig

GenerationsEvalListItem

GetEvalDraftResponse

GroupCall

GroupMessage

GroupResource

GroupRun

ImportField

InvocationSlot

ListEvalApiEval

ListFilterOption

ListFilterSection

MvInfo

OperationInfo

OperationPrompts

PageMetaItem

ParamInfo

ProducedMedia

ProfileSummary

QGetProfileContextV4RoleResource

RunStatus

TableInfo

ThemeBundle

ThemePrimitives

ThemeTokens

Thresholds

UpdateEvalItem

UpdateEvalPatch

app__infra__eval__types__DraftFormState

app__infra__eval__types__SectionFilter