🦋 New version release

typescript/core/CHANGELOG.md

@@ -1,5 +1,22 @@
 # @smooai/smooth-operator-core
 
+## 0.20.3
+
+### Patch Changes
+
+- ab24904: feat(llm,python): clamp `max_tokens` to the model's output ceiling (Python parity)
+
+  Python parity for the Rust `LlmClient` output-ceiling clamp. `AgentOptions` gains a
+  `model_max_output: int | None` field and the engine now sends
+  `effective_max_tokens(max_tokens, model_max_output)` =
+  `min(max_tokens, ceiling)` on both the streaming and non-streaming chat paths (`None`
+  / non-positive ceiling ⇒ graceful passthrough, no behaviour change). A new
+  `effective_max_tokens` helper is exported for consumers. This stops a policy/budget
+  `max_tokens` from exceeding what a model can physically emit — which otherwise makes
+  a reasoning model burn its budget and return empty, or 400s upstream (e.g.
+  `groq-compound`'s 8192 output cap). The ceiling is sourced from the gateway's
+  `/model/info` by the consumer (the server), kept out of the engine. EPIC th-1cc9fa.
+
 ## 0.20.2
 
 ### Patch Changes

typescript/core/package.json

@@ -1,6 +1,6 @@
 {
     "name": "@smooai/smooth-operator-core",
-    "version": "0.20.2",
+    "version": "0.20.3",
     "description": "Native TypeScript implementation of the smooth-operator agent engine — an in-process, OpenAI-compatible agentic tool-calling loop with knowledge grounding. The TypeScript sibling of the Rust reference engine, the C# core, and the Python core.",
     "type": "module",
     "license": "MIT",