fix: skip config env var processing for flows without Config parameters

When a flow with a deploy-time Config triggers another flow via
DeployedFlow, the parent flow's METAFLOW_FLOW_CONFIG_VALUE env var
leaked into the child flow's click API, causing either "Options were
not properly set" or an AttributeError on the False-valued
config_input. Guard _compute_flow_parameters so the config-resolution
block only runs when the flow actually declares Config parameters.

Fixes #2717

Author: Nissan Pow

metaflow/runner/click_api.py

@@ -460,7 +460,7 @@ def _compute_flow_parameters(self):
         self._cached_computed_parameters = []
 
         config_options = None
-        if CLICK_API_PROCESS_CONFIG:
+        if CLICK_API_PROCESS_CONFIG and self._config_input:
             with flow_context(self._flow_cls) as _:
                 # We are going to resolve the configs first and then get the parameters.
                 # Note that configs may update/add parameters so the order is important

test/unit/test_deployed_flow_config.py

@@ -0,0 +1,88 @@
+"""
+Regression test for GitHub issue #2717.
+
+When a flow with a deploy-time Config triggers another flow via DeployedFlow,
+the parent flow's METAFLOW_FLOW_CONFIG_VALUE environment variable was leaking
+into the child flow's click API initialization, causing:
+
+    "Options were not properly set -- this is an internal error."
+
+The fix ensures that _compute_flow_parameters skips config env var processing
+when the flow has no config parameters (self._config_input is False).
+"""
+
+import importlib
+import json
+import os
+import sys
+import tempfile
+
+import pytest
+
+
+@pytest.fixture
+def fake_flow_file(tmp_path):
+    """Create a temporary flow file without any Config parameters."""
+    flow_code = """\
+from metaflow import FlowSpec, Parameter, step
+
+class FakeTargetFlow(FlowSpec):
+    alpha = Parameter("alpha", type=str, help="A param", required=False)
+
+    @step
+    def start(self):
+        self.next(self.end)
+
+    @step
+    def end(self):
+        pass
+
+if __name__ == "__main__":
+    FakeTargetFlow()
+"""
+    flow_file = tmp_path / "fake_target_flow.py"
+    flow_file.write_text(flow_code)
+    return str(flow_file)
+
+
+def test_no_config_flow_ignores_config_env_var(fake_flow_file, monkeypatch):
+    """
+    A flow without Config parameters should not fail when
+    METAFLOW_FLOW_CONFIG_VALUE is set in the environment (e.g., from a parent
+    flow that does have Config).
+
+    This simulates the DeployedFlow trigger path: the parent flow sets
+    METAFLOW_FLOW_CONFIG_VALUE, then DeployedFlow creates a MetaflowAPI
+    from a fake flow file (no configs). Accessing a sub-command (like 'run')
+    triggers _compute_flow_parameters which must not choke on the stale
+    env var.
+    """
+    # Simulate the parent flow having set config env vars
+    monkeypatch.setenv(
+        "METAFLOW_FLOW_CONFIG_VALUE", json.dumps({"config": {"key": "value"}})
+    )
+    monkeypatch.setenv("METAFLOW_FLOW_CONFIG", json.dumps({"config": "kv.config"}))
+
+    from metaflow.parameters import flow_context
+
+    # Reload CLI modules to get a clean state (same as DeployerImpl does)
+    with flow_context(None):
+        for module_name in [
+            "metaflow.cli",
+            "metaflow.cli_components.run_cmds",
+            "metaflow.cli_components.init_cmd",
+        ]:
+            if module_name in sys.modules:
+                importlib.reload(sys.modules[module_name])
+
+    from metaflow.cli import start
+    from metaflow.runner.click_api import MetaflowAPI
+
+    api_func = MetaflowAPI.from_cli(fake_flow_file, start)
+    api = api_func()
+
+    # Accessing a sub-command triggers _compute_flow_parameters which
+    # previously raised "Options were not properly set" when config env
+    # vars from a parent flow were present but the target flow had none.
+    run_cmd = api.run
+    assert run_cmd is not None