Merge a4f021cfa0 into a81f507bca

litellm/proxy/guardrails/guardrail_hooks/custom_code/sandbox.py

@@ -31,6 +31,7 @@ from RestrictedPython.Eval import default_guarded_getitem
 from RestrictedPython.Guards import (
     full_write_guard,
     guarded_iter_unpack_sequence,
+    guarded_unpack_sequence,
     safer_getattr,
 )
 from RestrictedPython.transformer import copy_locations
@@ -46,7 +47,10 @@ class AsyncAwareTransformer(RestrictingNodeTransformer):
     has the same ``_fields`` as ``FunctionDef`` and the same security
     semantics, so we delegate to ``visit_FunctionDef`` — name check, argument
     check, print-scope wrapping, and any future additions to that method are
-    inherited automatically. ``AsyncFor``/``AsyncWith``/``Await`` delegate to
+    inherited automatically. ``AsyncWith`` gets the same treatment for the same
+    reason: ``node_contents_visit`` only recurses into children, so routing it
+    there left ``async with x as (a, b)`` without the unpack guard that
+    ``with x as (a, b)`` gets. ``AsyncFor``/``Await`` delegate to
     ``node_contents_visit`` so their children still get transformed.
 
     ``visit_While`` rewrites ``while test:`` to ``while _budget_ok_() and test:``
@@ -77,7 +81,7 @@ class AsyncAwareTransformer(RestrictingNodeTransformer):
         return self.node_contents_visit(node)
 
     def visit_AsyncWith(self, node: ast.AsyncWith) -> ast.AST:
-        return self.node_contents_visit(node)
+        return self.visit_With(node)
 
     def visit_Await(self, node: ast.Await) -> ast.AST:
         return self.node_contents_visit(node)
@@ -137,6 +141,12 @@ def build_sandbox_globals() -> dict[str, object]:
         "_getitem_": default_guarded_getitem,
         "_getiter_": budgeted_iter,
         "_iter_unpack_sequence_": guarded_iter_unpack_sequence,
+        # RestrictedPython emits _unpack_sequence_ for every tuple-unpacking
+        # target that is not a for-loop target -- ``a, b = pair``,
+        # ``a, *rest = seq``, ``with x as (a, b)`` -- and, like _inplacevar_,
+        # ships no default. Without it those statements compile and then raise
+        # NameError the first time the guardrail runs.
+        "_unpack_sequence_": guarded_unpack_sequence,
         "_write_": full_write_guard,
         "_inplacevar_": _inplacevar_,
         "_budget_ok_": budget_ok,

tests/unit/proxy/guardrails/test_custom_code_security.py

@@ -846,3 +846,86 @@ def test_config_without_timeout_uses_the_default():
         _initialize_from_config("custom-code-default-timeout", {}).execution_timeout
         == DEFAULT_EXECUTION_TIMEOUT_SECONDS
     )
+
+
+@pytest.mark.parametrize(
+    ("body", "expected"),
+    [
+        # Every one of these compiles fine and then raises
+        # "NameError: name '_unpack_sequence_' is not defined" at call time when
+        # the guard is missing from the sandbox globals.
+        ("    a, b = inputs['pair']\n    return a + b\n", 3),
+        ("    a, (b, c) = inputs['nested']\n    return a + b + c\n", 6),
+        ("    a, *rest = inputs['seq']\n    return rest\n", [2, 3]),
+        ("    with inputs['ctx'] as (a, b):\n        return a + b\n", 15),
+    ],
+)
+def test_tuple_unpacking_runs(body: str, expected):
+    """Ordinary tuple unpacking must work inside a guardrail, not NameError."""
+
+    class _Ctx:
+        def __enter__(self):
+            return (7, 8)
+
+        def __exit__(self, *args):
+            return False
+
+    guardrail = _compile("def apply_guardrail(inputs, request_data, input_type):\n" + body)
+    fn = guardrail._compiled_function
+    assert fn is not None
+    inputs = {"pair": (1, 2), "nested": (1, (2, 3)), "seq": [1, 2, 3], "ctx": _Ctx()}
+    assert fn(inputs, {}, "request") == expected
+
+
+def _guard_names(source: str) -> set[str]:
+    """Names of the RestrictedPython guards the compiled bytecode calls."""
+    import types
+
+    from litellm.proxy.guardrails.guardrail_hooks.custom_code.sandbox import (
+        compile_sandboxed,
+    )
+
+    found: set[str] = set()
+    stack = [compile_sandboxed(source)]
+    while stack:
+        code = stack.pop()
+        found |= {n for n in code.co_names + code.co_varnames if n.startswith("_") and n.endswith("_")}
+        stack += [c for c in code.co_consts if isinstance(c, types.CodeType)]
+    return found
+
+
+def test_async_with_gets_the_same_guards_as_with():
+    """`async with` is the async spelling of `with`; it must not enforce less."""
+    sync_guards = _guard_names("def f(x):\n    with x as (a, b):\n        pass\n")
+    async_guards = _guard_names("async def f(x):\n    async with x as (a, b):\n        pass\n")
+
+    assert "_unpack_sequence_" in sync_guards, "precondition: `with` unpacking is guarded"
+    assert sync_guards <= async_guards
+
+
+@pytest.mark.asyncio
+async def test_async_with_still_executes():
+    """Guarding `async with` must not break it."""
+    from litellm.proxy.guardrails.guardrail_hooks.custom_code.sandbox import (
+        build_sandbox_globals,
+        compile_sandboxed,
+    )
+
+    class _ACtx:
+        def __init__(self, value):
+            self.value = value
+
+        async def __aenter__(self):
+            return self.value
+
+        async def __aexit__(self, *args):
+            return False
+
+    sandbox_globals = build_sandbox_globals()
+    exec(  # noqa: S102
+        compile_sandboxed(
+            "async def f(ctx):\n    async with ctx as (a, b):\n        return a + b\n"
+        ),
+        sandbox_globals,
+    )
+    assert await sandbox_globals["f"](_ACtx((5, 6))) == 11