@mastra/schema-compat 1.3.3-alpha.0 → 1.3.4-alpha.0

package/CHANGELOG.md

@@ -1,5 +1,31 @@
 # @mastra/schema-compat
 
+## 1.3.4-alpha.0
+
+### Patch Changes
+
+- Fixed Meta (Llama) and DeepSeek schemas leaking raw number bounds to the model. A field like `z.number().int()` was sent to the model with bogus `minimum: -9007199254740991` / `maximum: 9007199254740991` values, and `z.number().min(1).max(50)` leaked `minimum`/`maximum` keywords, even though OpenAI, Google, and Anthropic already strip these. Numeric constraints are now moved into the field description for Meta and DeepSeek too, matching the other providers. ([#19073](https://github.com/mastra-ai/mastra/pull/19073))
+
+  **Before** (Meta/DeepSeek, `z.object({ age: z.number().min(0).max(120) })`):
+
+  ```json
+  { "age": { "type": "number", "minimum": 0, "maximum": 120 } }
+  ```
+
+  **After**:
+
+  ```json
+  { "age": { "type": "number", "description": "constraints: greater than or equal to 0, lower than or equal to 120" } }
+  ```
+
+  Closes #19072.
+
+## 1.3.3
+
+### Patch Changes
+
+- Fix the Zod v4 nullable and optional handlers gating on the wrapper type instead of the wrapped inner type. They checked `value.constructor.name` (always `"ZodNullable"`/`"ZodOptional"`), so the inner type was always processed. A nullable/optional wrapping an unsupported inner type (such as a tuple) is now passed through unchanged, matching the v3 handler, instead of being processed and rejected. Closes #18687. ([#18688](https://github.com/mastra-ai/mastra/pull/18688))
+
 ## 1.3.3-alpha.0
 
 ### Patch Changes