Add provider for-tools command and max_tokens support

- Add 'cmdforge providers for-tools <tool>...' to list providers used by tools
- Add --warm flag to pre-load local models (Ollama) for faster first inference
- Warm larger models first to claim contiguous GPU memory
- Skip cloud providers (claude, opencode, gemini) during warm-up
- Add max_tokens field to PromptStep for controlling output length
- Pass max_tokens to call_provider with provider-specific flags
- Add comprehensive tests for provider functionality

Co-Authored-By: Claude Opus 4.5 <[email protected]>

src/cmdforge/runner.py

@@ -546,7 +546,7 @@ def execute_prompt_step(
         if provider.lower() == "mock":
             result = mock_provider(prompt)
         else:
-            result = call_provider(provider, prompt)
+            result = call_provider(provider, prompt, max_tokens=step.max_tokens)
 
         if not result.success:
             print(f"Error in prompt step: {result.error}", file=sys.stderr)
@@ -590,7 +590,7 @@ Please try again with valid JSON matching the schema exactly."""
         if provider.lower() == "mock":
             result = mock_provider(current_prompt)
         else:
-            result = call_provider(provider, current_prompt)
+            result = call_provider(provider, current_prompt, max_tokens=step.max_tokens)
 
         if not result.success:
             print(f"Error in prompt step: {result.error}", file=sys.stderr)