Add go-llm v2: redesigned API for simpler LLM abstraction

v2 is a new Go module (v2/) with a dramatically simpler API:
- Unified Message type (no more Input marker interface)
- Define[T] for ergonomic tool creation with standard context.Context
- Chat session with automatic tool-call loop (agent loop)
- Streaming via pull-based StreamReader
- MCP one-call connect (MCPStdioServer, MCPHTTPServer, MCPSSEServer)
- Middleware support (logging, retry, timeout, usage tracking)
- Decoupled JSON Schema (map[string]any, no provider coupling)
- Sample tools: WebSearch, Browser, Exec, ReadFile, WriteFile, HTTP
- Providers: OpenAI, Anthropic, Google (all with streaming)

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

v2/provider/provider.go

@@ -0,0 +1,92 @@
+// Package provider defines the interface that LLM backend implementations must satisfy.
+package provider
+
+import "context"
+
+// Message is the provider-level message representation.
+type Message struct {
+	Role       string
+	Content    string
+	Images     []Image
+	ToolCalls  []ToolCall
+	ToolCallID string
+}
+
+// Image represents an image attachment at the provider level.
+type Image struct {
+	URL         string
+	Base64      string
+	ContentType string
+}
+
+// ToolCall represents a tool invocation requested by the model.
+type ToolCall struct {
+	ID        string
+	Name      string
+	Arguments string // raw JSON
+}
+
+// ToolDef defines a tool available to the model.
+type ToolDef struct {
+	Name        string
+	Description string
+	Schema      map[string]any // JSON Schema
+}
+
+// Request is a completion request at the provider level.
+type Request struct {
+	Model       string
+	Messages    []Message
+	Tools       []ToolDef
+	Temperature *float64
+	MaxTokens   *int
+	TopP        *float64
+	Stop        []string
+}
+
+// Response is a completion response at the provider level.
+type Response struct {
+	Text      string
+	ToolCalls []ToolCall
+	Usage     *Usage
+}
+
+// Usage captures token consumption.
+type Usage struct {
+	InputTokens  int
+	OutputTokens int
+	TotalTokens  int
+}
+
+// StreamEventType identifies the kind of stream event.
+type StreamEventType int
+
+const (
+	StreamEventText      StreamEventType = iota // Text content delta
+	StreamEventToolStart                        // Tool call begins
+	StreamEventToolDelta                        // Tool call argument delta
+	StreamEventToolEnd                          // Tool call complete
+	StreamEventDone                             // Stream complete
+	StreamEventError                            // Error occurred
+)
+
+// StreamEvent represents a single event in a streaming response.
+type StreamEvent struct {
+	Type      StreamEventType
+	Text      string
+	ToolCall  *ToolCall
+	ToolIndex int
+	Error     error
+	Response  *Response
+}
+
+// Provider is the interface that LLM backends implement.
+type Provider interface {
+	// Complete performs a non-streaming completion.
+	Complete(ctx context.Context, req Request) (Response, error)
+
+	// Stream performs a streaming completion, sending events to the channel.
+	// The provider MUST close the channel when done.
+	// The provider MUST send exactly one StreamEventDone as the last non-error event.
+	Stream(ctx context.Context, req Request, events chan<- StreamEvent) error
+}