feat: add conversational display mode types

[shannonsuhendra]

No description provided.

[claude]

Per the conventions in agent_docs/conventions.md: "Use enums for fixed value sets — NEVER leave raw strings/numbers. Raw values lose type safety and autocomplete."

This should be an enum, not a string union:

Suggested change

	export type ConversationalDisplayMode = 'FullTrace' | 'InputsAndOutputs' | 'ToolNameOnly';
	export enum ConversationalDisplayMode {
	FullTrace = 'FullTrace',
	InputsAndOutputs = 'InputsAndOutputs',
	ToolNameOnly = 'ToolNameOnly',
	}

[claude]

All other fields in AgentAppearance are optional (welcomeTitle?, welcomeDescription?, startingPrompts?), and appearance itself is optional on RawAgentGetByIdResponse. If the API can return appearance without displayMode (e.g., for agents created before this field was introduced), marking it required causes runtime undefined access.

Per conventions: "Mark optional fields as optional in type interfaces — over-requiring causes runtime undefined access on fields the API didn't return."

Unless the API guarantees displayMode is always present whenever appearance is returned, this should be:

Suggested change

	displayMode: ConversationalDisplayMode;
	displayMode?: ConversationalDisplayMode;

[claude]

Review findings

Two issues in src/models/conversational-agent/agents/agents.types.ts:

1. Line 22 — String union instead of enum: ConversationalDisplayMode uses raw string literals. Convention requires an enum for fixed value sets.

2. Line 35 — Potentially over-required field: displayMode is non-optional while every other field in AgentAppearance is optional. If the API can omit it, callers will get runtime undefined access.