fix(router/discovery): always probe ollama capabilities, cache is optional

DiscoverOllama() interpreted a nil probeCache as 'skip probing
entirely' rather than 'probe but don't cache.' cmd/gnoma/main.go's
synchronous discovery path passes nil, so every ollama-discovered
model got SupportsTools=false (the Go zero value), regardless of
what ollama actually reported in its capabilities field.

The symptom: filterFeasible rejected every ollama arm for any
tool-requiring task with reason=tools_required_but_unsupported,
even when ollama itself reported the model as tool-capable. Verified
via curl: qwen3:14b advertises capabilities=[completion, tools,
thinking] and has 'tools' in its template, but the gnoma arm shipped
with tool_use_capability=false.

Fix: always run probeOllamaModel; treat probeCache as an optional
memoisation aid only. nil cache now means 'no caching across calls'
not 'no probing.' For users with many models, passing a real cache
still avoids redundant HTTP calls — semantics for that path are
unchanged.

Surfaced via the new filterFeasible Debug logging from the previous
commit, which made the per-arm rejection reasons visible.

internal/router/discovery.go

@@ -93,16 +93,27 @@ func DiscoverOllama(ctx context.Context, baseURL string, probeCache map[string]O
 			Size:     m.Size,
 		}
 
+		// Always probe; the cache is optional. Previously nil-cache was
+		// treated as "skip probing entirely", which left SupportsTools
+		// at its zero value (false) for every model — every ollama-
+		// discovered arm then got marked as tool-unsupported and
+		// rejected by filterFeasible for any tool-requiring task. main.go
+		// passes nil from the synchronous discovery path; we still want
+		// real probe data there.
+		var result OllamaProbeResult
 		if probeCache != nil {
-			result, ok := probeCache[m.Name]
-			if !ok {
+			if cached, ok := probeCache[m.Name]; ok {
+				result = cached
+			} else {
 				result = probeOllamaModel(ctx, baseURL, m.Name)
 				probeCache[m.Name] = result
 			}
-			dm.SupportsTools = result.SupportsTools
-			dm.SupportsVision = result.SupportsVision
-			dm.ContextSize = result.ContextSize
+		} else {
+			result = probeOllamaModel(ctx, baseURL, m.Name)
 		}
+		dm.SupportsTools = result.SupportsTools
+		dm.SupportsVision = result.SupportsVision
+		dm.ContextSize = result.ContextSize
 
 		if dm.ContextSize == 0 {
 			dm.ContextSize = defaultOllamaContextSize