feat: add OLLAMA_NUM_CTX config to cap context window (#83)

- Add ollama_num_ctx setting (default 4096) to config.py
- Pass num_ctx option to Ollama in agent.py and agents/base.py
- Add OLLAMA_NUM_CTX to .env.example with usage docs
- Add context_window note in providers.yaml
- Fix mock_settings in test_agent.py for new attribute
- qwen3:30b with 4096 ctx uses ~19GB vs 45GB default

.env.example

@@ -17,6 +17,11 @@
 # LLM model to use via Ollama (default: qwen3.5:latest)
 # OLLAMA_MODEL=qwen3.5:latest
 
+# Ollama context window size (default: 4096 tokens)
+# Set higher for more context, lower to save RAM. 0 = model default.
+# qwen3:30b + 4096 ctx ≈ 19GB VRAM; default ctx ≈ 45GB.
+# OLLAMA_NUM_CTX=4096
+
 # Enable FastAPI interactive docs at /docs and /redoc (default: false)
 # DEBUG=true