added documentation

Author: scalabreseGD

docs/ref/extensions/sandbox/litellm_compaction.md

@@ -0,0 +1,6 @@
+# `LiteLLM Compaction`
+
+::: agents.extensions.sandbox.litellm_compaction
+    options:
+        members:
+            - LiteLLMCompaction

docs/sandbox/guide.md

@@ -202,6 +202,8 @@ Built-in capabilities include:
 
 By default, `SandboxAgent.capabilities` uses `Capabilities.default()`, which includes `Filesystem()`, `Shell()`, and `Compaction()`. If you pass `capabilities=[...]`, that list replaces the default, so include any default capabilities you still want.
 
+`Compaction` sizes its dynamic policy from a built-in OpenAI-only context-window registry. If you route requests through litellm to a non-OpenAI model (Anthropic, Bedrock, Vertex, custom proxy aliases, ...), the lookup misses and the policy falls back to a static 240k threshold regardless of the model's real input window. For those deployments use [`LiteLLMCompaction`][agents.extensions.sandbox.litellm_compaction.LiteLLMCompaction] instead — it resolves the cap through `litellm.get_model_info()` and exposes `LiteLLMCompaction.for_model(...)` and `LiteLLMCompaction.for_context_window(...)` factories. Requires the `openai-agents[litellm]` extra.
+
 For skills, choose the source based on how you want them materialized:
 
 - `Skills(lazy_from=LocalDirLazySkillSource(...))` is a good default for larger local skill directories because the model can discover the index first and load only what it needs.

mkdocs.yml

@@ -186,6 +186,8 @@ plugins:
                         - Any-LLM provider: ref/extensions/models/any_llm_provider.md
                         - LiteLLM model: ref/extensions/models/litellm_model.md
                         - LiteLLM provider: ref/extensions/models/litellm_provider.md
+                    - Sandbox capabilities:
+                        - LiteLLM compaction: ref/extensions/sandbox/litellm_compaction.md
                     - Tool output trimmer: ref/extensions/tool_output_trimmer.md
                     - SQLAlchemySession: ref/extensions/memory/sqlalchemy_session.md
                     - Async SQLite session: ref/extensions/memory/async_sqlite_session.md