Document --model-config / -m CLI flag in README

- Add -m to architecture diagram CLI flags listing
- Document model config persistence in session recovery section
- Add CLI override subsection under Model configs with usage example

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Author: anticomputer

README.md

@@ -24,7 +24,7 @@ You can find a detailed overview of the taskflow grammar [here](doc/GRAMMAR.md)
 ```
 ┌─────────────────────────────────────────────────────┐
 │                   CLI (cli.py)                      │
-│  Typer-based entry point: -p, -t, -l, -g, --resume  │
+│  Typer-based entry point: -p, -t, -l, -g, -m, --resume│
 └─────────────────────┬───────────────────────────────┘
                       │
 ┌─────────────────────▼───────────────────────────────┐
@@ -132,6 +132,14 @@ Resume from the last successful checkpoint:
 python -m seclab_taskflow_agent --resume abc123def456
 ```
 
+The session checkpoint persists the CLI-provided `--model-config` value (if
+any), so resumes use the same model configuration by default. To override the
+model config on resume, pass `--model-config` / `-m` explicitly:
+
+```bash
+python -m seclab_taskflow_agent --resume abc123def456 -m examples.model_configs.responses_api
+```
+
 Failed tasks are automatically retried up to 3 times with increasing backoff
 before the session is saved. Session checkpoints are stored in the
 platform-specific application data directory.
@@ -611,6 +619,19 @@ taskflow:
 
 The model version can then be updated by changing `gpt_latest` in the `model_config` file and applied across all taskflows that use the config.
 
+#### CLI override
+
+The `model_config` can also be specified (or overridden) from the command line
+with `--model-config` / `-m`. This takes precedence over any `model_config`
+value defined in the taskflow YAML:
+
+```bash
+python -m seclab_taskflow_agent -t examples.taskflows.echo -m examples.model_configs.responses_api
+```
+
+The CLI-provided model config is persisted in the session checkpoint, so
+resumed runs automatically use the same configuration.
+
 In addition, model specific parameters can be provided via `model_config`. To do so, define a `model_settings` section in the `model_config` file. This section has to be a dictionary with the model names as keys:
 
 ```yaml

tests/test_runner.py

@@ -286,26 +286,24 @@ class TestCliModelConfigOverride:
 
     def test_cli_overrides_taskflow_model_config(self):
         """cli_model_config takes precedence over taskflow_doc.model_config_ref."""
-        taskflow_ref = "taskflow.models.default"
-        cli_ref = "cli.models.override"
-
-        # Simulate the override logic from run_main
-        model_config_ref = taskflow_ref
-        if cli_ref:
-            model_config_ref = cli_ref
-
-        assert model_config_ref == cli_ref
+        model_config_ref = self._resolve("taskflow.models.default", "cli.models.override")
+        assert model_config_ref == "cli.models.override"
 
     def test_taskflow_model_config_used_when_cli_absent(self):
         """Taskflow model_config_ref is used when cli_model_config is None."""
         taskflow_ref = "taskflow.models.default"
         cli_ref = None
 
+        model_config_ref = self._resolve(taskflow_ref, cli_ref)
+        assert model_config_ref == taskflow_ref
+
+    @staticmethod
+    def _resolve(taskflow_ref: str, cli_ref: str | None) -> str:
+        """Reproduce the override logic from run_main."""
         model_config_ref = taskflow_ref
         if cli_ref:
             model_config_ref = cli_ref
-
-        assert model_config_ref == taskflow_ref
+        return model_config_ref
 
     def test_cli_model_config_resolves_via_available_tools(self):
         """CLI-provided model config is resolved through _resolve_model_config."""