feat: import agents_and_robots platform as unibots (Matrix-out, unibus transport)

Reemplaza el scaffold del echobot por la plataforma completa de bots traida
desde ~/DataProyects/Github/agents_and_robots tras la operacion Matrix-out:
los bots ya no hablan por Matrix sino por el bus unibus (modelo todo-rooms +
E2E via shell/transportunibus sobre github.com/enmanuel/unibus/pkg/client).

- go.mod: replace de unibus -> ../unibus y de fn-registry -> ../../../.. (paths
  relativos reajustados a la nueva ubicacion dentro de fn_registry).
- app.md: bump a 0.2.0, descripcion + arquitectura + comandos + gotchas reales.
- modulo Go conservado como github.com/enmanuel/agents (sin reescribir imports).

agents_and_robots queda archivado como museo de la era Matrix.

shell/llm/factory.go

@@ -0,0 +1,51 @@
+package llm
+
+import (
+	"context"
+	"fmt"
+	"log/slog"
+
+	"github.com/enmanuel/agents/internal/config"
+	coretypes "github.com/enmanuel/agents/pkg/llm"
+)
+
+// FromConfig builds a CompleteFunc from an LLMProviderCfg.
+func FromConfig(cfg config.LLMProviderCfg, log *slog.Logger) (coretypes.CompleteFunc, error) {
+	log.Info("llm_provider_init", "provider", cfg.Provider, "model", cfg.Model)
+	switch cfg.Provider {
+	case "anthropic":
+		return NewAnthropicComplete(cfg.APIKeyEnv, cfg.BaseURL, log), nil
+	case "openai":
+		return NewOpenAIComplete(cfg.APIKeyEnv, cfg.BaseURL, log), nil
+	case "ollama":
+		base := cfg.BaseURL
+		if base == "" {
+			base = "http://localhost:11434/v1"
+		}
+		return NewOpenAIComplete("OLLAMA_API_KEY", base, log), nil
+	case "claude-code":
+		return NewClaudeCodeComplete(cfg.ClaudeCode, log), nil
+	default:
+		return nil, fmt.Errorf("unknown LLM provider: %s", cfg.Provider)
+	}
+}
+
+// WithFallback wraps primary with a fallback CompleteFunc.
+// If primary returns an error, fallback is tried with the fallback config's model.
+func WithFallback(primary, fallback coretypes.CompleteFunc, fallbackCfg config.LLMProviderCfg, log *slog.Logger) coretypes.CompleteFunc {
+	return func(ctx context.Context, req coretypes.CompletionRequest) (coretypes.CompletionResponse, error) {
+		resp, err := primary(ctx, req)
+		if err != nil {
+			log.Warn("llm_fallback_triggered", "primary_err", err)
+			// Override request fields with fallback config values
+			if fallbackCfg.Model != "" {
+				req.Model = fallbackCfg.Model
+			}
+			if fallbackCfg.MaxTokens > 0 {
+				req.MaxTokens = fallbackCfg.MaxTokens
+			}
+			return fallback(ctx, req)
+		}
+		return resp, nil
+	}
+}