v1.6 System Prompt 工程化 + 渐进式技能发现

重构 system prompt 为动态组装系统：PromptComposer 聚合核心身份、AGENTS.md
项目规范和 Skills 目录；新增 read_skill 工具实现技能按需加载；包重命名
internal/context -> internal/prompt 消除 stdlib 冲突。

internal/engine/loop.go

@@ -6,6 +6,7 @@ import (
 	"log"
 	"sync"
 
+	"go-tiny-claw/internal/prompt"
 	"go-tiny-claw/internal/provider"
 	"go-tiny-claw/internal/schema"
 	"go-tiny-claw/internal/tools"
@@ -19,6 +20,7 @@ type AgentEngine struct {
 	// WorkDir (工作区): 借鉴 OpenClaw 的理念，Agent 必须有一个明确的物理边界
 	WorkDir        string
 	EnableThinking bool // 【新增】慢思考模式开关
+	composer       *prompt.PromptComposer
 }
 
 func NewAgentEngine(p provider.LLMProvider, r tools.Registry, workDir string, enableThinking bool) *AgentEngine {
@@ -27,6 +29,7 @@ func NewAgentEngine(p provider.LLMProvider, r tools.Registry, workDir string, en
 		registry:       r,
 		WorkDir:        workDir,
 		EnableThinking: enableThinking,
+		composer:       prompt.NewPromptComposer(workDir),
 	}
 }
 
@@ -60,15 +63,11 @@ func (e *AgentEngine) Run(ctx context.Context, userPrompt string) error {
 	log.Printf("[Engine] 引擎启动，锁定工作区: %s\n", e.WorkDir)
 	log.Printf("[Engine] 慢思考模式 (Thinking Phase): %v\n", e.EnableThinking)
 
+	systemMsg := e.composer.Build()
+
 	contextHistory := []schema.Message{
-		{
-			Role:    schema.RoleSystem,
-			Content: "You are go-tiny-claw, an expert coding assistant. You have full access to tools in the workspace.",
-		},
-		{
-			Role:    schema.RoleUser,
-			Content: userPrompt,
-		},
+		systemMsg, // 注入动态组装的内核、AGENTS.md 与 Skills
+		{Role: schema.RoleUser, Content: userPrompt},
 	}
 
 	turnCount := 0