fix(ai,agent,coding-agent): normalize null message content at ingestion boundaries (#6343)

The Message types require content to always be present, but untyped JS
extension tools, hand-built histories, and old or hand-edited session
files can violate that contract, crashing rendering, compaction, and
provider request conversion with 'content is not iterable'.

Normalize null/missing content to an empty array at the ingestion
boundaries instead of guarding every consumer:

- transformMessages (choke point before every provider request)
- createToolResultMessage in the agent loop
- session entry loading (message and custom_message entries)
- extension custom messages (sendCustomMessage, before_agent_start)
- message_end extension replacements

fixes #6259, fixes #6276

packages/ai/src/api/transform-messages.ts

@@ -68,7 +68,10 @@ export function transformMessages<TApi extends Api>(
 ): Message[] {
 	// Build a map of original tool call IDs to normalized IDs
 	const toolCallIdMap = new Map<string, string>();
-	const imageAwareMessages = downgradeUnsupportedImages(messages, model);
+	// Normalize null/undefined content from untyped callers (custom tools, hand-built
+	// histories, old session files) so downstream code can rely on the type contract.
+	const normalizedMessages = messages.map((msg) => (msg.content == null ? { ...msg, content: [] } : msg));
+	const imageAwareMessages = downgradeUnsupportedImages(normalizedMessages, model);
 
 	// First pass: transform messages (unsupported image downgrade, thinking blocks, tool call ID normalization)
 	const transformed = imageAwareMessages.map((msg) => {