package main import ( "bytes" "context" "encoding/base64" "encoding/json" "fmt" "io" "log" "net" "net/http" "net/url" "os" "os/exec" "regexp" "strconv" "strings" "sync" "time" "unicode" sdk "gitcode.com/JianFeeeee/homeagent-sdk/sdk" "github.com/chromedp/chromedp" ) type Plugin struct { name string sdk *sdk.PluginSDK mu sync.RWMutex timeout int proxy string dataDir string client *http.Client sessions map[string]*BrowserSession nextID int wg sync.WaitGroup stopCh chan struct{} } type BrowserSession struct { id string allocCtx context.Context cancel context.CancelFunc ctx context.Context createdAt time.Time timeout time.Duration closed bool mu sync.Mutex currentURL string } func NewPlugin(name string, config map[string]interface{}) (sdk.Plugin, error) { return &Plugin{name: name, stopCh: make(chan struct{}), sessions: make(map[string]*BrowserSession)}, nil } func (p *Plugin) Name() string { return p.name } func readCfg[T string | int64 | float64](s sdk.SettingsAPI, key string, fallback T) T { v, err := s.Get(key) if err == nil && v != nil { if s, ok := v.(string); ok && s != "" { switch any(fallback).(type) { case string: return any(s).(T) case int64: if n, err := strconv.ParseInt(s, 10, 64); err == nil { return any(n).(T) } case float64: if n, err := strconv.ParseFloat(s, 64); err == nil { return any(n).(T) } } } } v2, err2 := s.GetCore("plugin." + "browser" + "." + key) if err2 == nil && v2 != nil { if s, ok := v2.(string); ok && s != "" { switch any(fallback).(type) { case string: return any(s).(T) case int64: if n, err := strconv.ParseInt(s, 10, 64); err == nil { return any(n).(T) } case float64: if n, err := strconv.ParseFloat(s, 64); err == nil { return any(n).(T) } } } } return fallback } func readArg[T string | int64 | float64](args map[string]interface{}, key string, fallback T) T { v, ok := args[key] if !ok || v == nil { return fallback } switch any(fallback).(type) { case string: if s, ok := v.(string); ok { return any(s).(T) } case int64: switch val := v.(type) { case float64: return any(int64(val)).(T) case string: if n, err := strconv.ParseInt(val, 10, 64); err == nil { return any(n).(T) } } case float64: switch val := v.(type) { case float64: return any(val).(T) case string: if n, err := strconv.ParseFloat(val, 64); err == nil { return any(n).(T) } } } return fallback } func errResult(msg string) map[string]interface{} { return map[string]interface{}{"isError": true, "content": msg} } func newHTTPClient(timeout int, proxyURL string) *http.Client { transport := &http.Transport{ DialContext: (&net.Dialer{ Timeout: time.Duration(timeout) * time.Second, KeepAlive: 30 * time.Second, }).DialContext, TLSHandshakeTimeout: time.Duration(timeout) * time.Second, ResponseHeaderTimeout: time.Duration(timeout) * time.Second, } if proxyURL != "" { if u, err := url.Parse(proxyURL); err == nil { transport.Proxy = http.ProxyURL(u) } } return &http.Client{ Timeout: time.Duration(timeout) * time.Second, Transport: transport, CheckRedirect: func(req *http.Request, via []*http.Request) error { if len(via) >= 5 { return fmt.Errorf("too many redirects") } return nil }, } } // ── Start / Stop ────────────────────────────────────────── func (p *Plugin) Start(s *sdk.PluginSDK) error { p.sdk = s s.SetAutoRestart(true) s.Settings().RegisterDef(sdk.ConfigDef{ Key: "timeout", Default: "30", Type: "int", DisplayName: "HTTP 超时(秒)", Description: "HTTP 请求超时时间", Category: "browser", }) s.Settings().RegisterDef(sdk.ConfigDef{ Key: "proxy", Default: "", Type: "string", DisplayName: "HTTP 代理", Description: "HTTP 代理地址,如 http://proxy:port", Category: "browser", }) s.Settings().RegisterDef(sdk.ConfigDef{ Key: "data_dir", Default: "", Type: "string", DisplayName: "浏览器数据目录", Description: "Chromium 用户数据目录路径(持久化 cookies/登录状态)。留空则每次启动临时目录。", Category: "browser", }) t := readCfg(s.Settings(), "timeout", float64(30)) p.timeout = int(t) if p.timeout < 5 { p.timeout = 5 } if p.timeout > 120 { p.timeout = 120 } p.proxy = readCfg(s.Settings(), "proxy", "") p.dataDir = readCfg(s.Settings(), "data_dir", "") if p.dataDir != "" { os.MkdirAll(p.dataDir, 0700) } p.client = newHTTPClient(p.timeout, p.proxy) tp := p.name + "_" s.RegisterTool(tp+"search", sdk.ToolDef{ Name: tp + "search", Description: "使用 Bing 搜索网页。返回标题、URL 和摘要。", Parameters: map[string]interface{}{ "type": "object", "properties": map[string]interface{}{ "query": map[string]interface{}{"type": "string", "description": "搜索关键词"}, "count": map[string]interface{}{"type": "integer", "description": "结果数量(1-20,默认5)"}, }, "required": []string{"query"}, }, }, p.handleSearch) s.RegisterTool(tp+"fetch", sdk.ToolDef{ Name: tp + "fetch", Description: "快速抓取 URL 内容(quick 模式)。纯 HTTP 请求,不支持 JS 渲染,有 SSRF 防护。", Parameters: map[string]interface{}{ "type": "object", "properties": map[string]interface{}{ "url": map[string]interface{}{"type": "string", "description": "HTTP/HTTPS URL"}, "max_chars": map[string]interface{}{"type": "integer", "description": "最大返回字符数(默认20000)"}, }, "required": []string{"url"}, }, }, p.handleFetch) s.RegisterTool(tp+"render", sdk.ToolDef{ Name: tp + "render", Description: "无头 Chromium 渲染网页并提取文本(normal 模式)。支持 JS 渲染的页面。", Parameters: map[string]interface{}{ "type": "object", "properties": map[string]interface{}{ "url": map[string]interface{}{"type": "string", "description": "URL"}, "wait": map[string]interface{}{"type": "integer", "description": "等待 JS 渲染的秒数(默认0)"}, }, "required": []string{"url"}, }, }, p.handleRender) s.RegisterTool(tp+"start", sdk.ToolDef{ Name: tp + "start", Description: "启动交互式浏览器会话(interactive 模式)。通过 CDP 连接 Chromium,支持导航、截图、点击、输入等操作。返回会话 ID。", Parameters: map[string]interface{}{ "type": "object", "properties": map[string]interface{}{ "url": map[string]interface{}{"type": "string", "description": "初始导航 URL(可选)"}, "timeout": map[string]interface{}{"type": "string", "description": "会话超时(如 5m, 10m,默认 10m)"}, }, }, }, p.handleBrowserStart) s.RegisterTool(tp+"navigate", sdk.ToolDef{ Name: tp + "navigate", Description: "在交互式浏览器中导航到指定 URL。自动等待页面 body 加载完成。", Parameters: map[string]interface{}{ "type": "object", "properties": map[string]interface{}{ "id": map[string]interface{}{"type": "string", "description": "浏览器会话 ID"}, "url": map[string]interface{}{"type": "string", "description": "目标 URL"}, "wait": map[string]interface{}{"type": "integer", "description": "页面加载后额外等待秒数(默认2,反爬页面建议5)"}, }, "required": []string{"id", "url"}, }, }, p.handleNavigate) s.RegisterTool(tp+"screenshot", sdk.ToolDef{ Name: tp + "screenshot", Description: "对交互式浏览器当前页面截图。返回 base64 编码的 PNG 图片。", Parameters: map[string]interface{}{ "type": "object", "properties": map[string]interface{}{ "id": map[string]interface{}{"type": "string", "description": "浏览器会话 ID"}, "full": map[string]interface{}{"type": "boolean", "description": "是否全页截图(默认 false,仅视口)"}, "format": map[string]interface{}{"type": "string", "description": "图片格式: png 或 jpeg(默认 png)"}, }, "required": []string{"id"}, }, }, p.handleScreenshot) s.RegisterTool(tp+"html", sdk.ToolDef{ Name: tp + "html", Description: "获取交互式浏览器当前页面 JS 渲染后的完整 HTML。用于模型分析页面结构、定位元素。", Parameters: map[string]interface{}{ "type": "object", "properties": map[string]interface{}{ "id": map[string]interface{}{"type": "string", "description": "浏览器会话 ID"}, "max_chars": map[string]interface{}{"type": "integer", "description": "最大返回字符数(默认50000)"}, }, "required": []string{"id"}, }, }, p.handleHTML) s.RegisterTool(tp+"click", sdk.ToolDef{ Name: tp + "click", Description: "在交互式浏览器中点击元素。自动等待元素可见后再点击。", Parameters: map[string]interface{}{ "type": "object", "properties": map[string]interface{}{ "id": map[string]interface{}{"type": "string", "description": "浏览器会话 ID"}, "selector": map[string]interface{}{"type": "string", "description": "CSS 选择器"}, "wait": map[string]interface{}{"type": "integer", "description": "等待元素出现的超时毫秒数(默认3000)"}, }, "required": []string{"id", "selector"}, }, }, p.handleClick) s.RegisterTool(tp+"type", sdk.ToolDef{ Name: tp + "type", Description: "在交互式浏览器中向输入框输入文字。自动等待元素可见、聚焦后清空再输入。", Parameters: map[string]interface{}{ "type": "object", "properties": map[string]interface{}{ "id": map[string]interface{}{"type": "string", "description": "浏览器会话 ID"}, "selector": map[string]interface{}{"type": "string", "description": "CSS 选择器"}, "text": map[string]interface{}{"type": "string", "description": "要输入的文字"}, "submit": map[string]interface{}{"type": "boolean", "description": "输入后按回车(默认 false)"}, "wait": map[string]interface{}{"type": "integer", "description": "等待元素出现的超时毫秒数(默认3000)"}, }, "required": []string{"id", "selector", "text"}, }, }, p.handleType) s.RegisterTool(tp+"scroll", sdk.ToolDef{ Name: tp + "scroll", Description: "在交互式浏览器中滚动页面。", Parameters: map[string]interface{}{ "type": "object", "properties": map[string]interface{}{ "id": map[string]interface{}{"type": "string", "description": "浏览器会话 ID"}, "dir": map[string]interface{}{"type": "string", "description": "方向: up, down, left, right(默认 down)"}, "amount": map[string]interface{}{"type": "integer", "description": "滚动像素数(默认 500)"}, }, "required": []string{"id"}, }, }, p.handleScroll) s.RegisterTool(tp+"close", sdk.ToolDef{ Name: tp + "close", Description: "关闭交互式浏览器会话,释放资源。", Parameters: map[string]interface{}{ "type": "object", "properties": map[string]interface{}{ "id": map[string]interface{}{"type": "string", "description": "浏览器会话 ID"}, }, "required": []string{"id"}, }, }, p.handleBrowserClose) p.wg.Add(1) go p.cleanupLoop() log.Printf("[%s] started, timeout=%ds proxy=%q", p.name, p.timeout, p.proxy) return nil } func (p *Plugin) Stop() error { close(p.stopCh) p.wg.Wait() if p.client != nil { p.client.CloseIdleConnections() } p.mu.Lock() for _, s := range p.sessions { s.Close() } p.sessions = nil p.mu.Unlock() log.Printf("[%s] stopped", p.name) return nil } // ── SSRF ────────────────────────────────────────────────── var privateCIDRs []*net.IPNet func init() { for _, c := range []string{ "127.0.0.0/8", "10.0.0.0/8", "172.16.0.0/12", "192.168.0.0/16", "100.64.0.0/10", "169.254.0.0/16", "::1/128", "fc00::/7", "fe80::/10", } { _, n, _ := net.ParseCIDR(c) if n != nil { privateCIDRs = append(privateCIDRs, n) } } } func isPrivateIP(ip net.IP) bool { for _, n := range privateCIDRs { if n.Contains(ip) { return true } } return false } func (p *Plugin) ssrfCheck(rawURL string) error { u, err := url.Parse(rawURL) if err != nil { return fmt.Errorf("invalid URL: %w", err) } if u.Scheme != "http" && u.Scheme != "https" { return fmt.Errorf("only http/https allowed, got: %s", u.Scheme) } ips, err := net.LookupHost(u.Hostname()) if err != nil { return fmt.Errorf("DNS lookup failed: %w", err) } for _, ip := range ips { if parsed := net.ParseIP(ip); parsed != nil && isPrivateIP(parsed) { return fmt.Errorf("blocked request to private IP: %s (%s)", u.Hostname(), ip) } } return nil } // ── Bing Search ─────────────────────────────────────────── type searchResult struct { Title, URL, Snippet string } func (p *Plugin) bingSearch(query string, count int) ([]searchResult, error) { u := fmt.Sprintf("https://www.bing.com/search?q=%s&count=%d", url.QueryEscape(query), count) req, _ := http.NewRequest("GET", u, nil) req.Header.Set("User-Agent", "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36") req.Header.Set("Accept-Language", "zh-CN,zh;q=0.9,en;q=0.8") resp, err := p.client.Do(req) if err != nil { return nil, fmt.Errorf("request failed: %w", err) } defer resp.Body.Close() body, _ := io.ReadAll(resp.Body) return parseBingResults(string(body), count), nil } func parseBingResults(html string, count int) []searchResult { var results []searchResult re := regexp.MustCompile(`
(.*?)
`) if sm := snipRe.FindStringSubmatch(block); len(sm) > 1 { r.Snippet = stripTags(sm[1]) } if r.URL != "" && r.Title != "" { results = append(results, r) } } return results } func (p *Plugin) handleSearch(args map[string]interface{}) (interface{}, error) { query := readArg(args, "query", "") if query == "" { return errResult("query is required"), nil } count := 5 if v, ok := args["count"].(float64); ok && v > 0 { count = int(v) } if count < 1 { count = 1 } if count > 20 { count = 20 } results, err := p.bingSearch(query, count) if err != nil { return errResult("search failed: " + err.Error()), nil } if len(results) == 0 { return map[string]interface{}{"content": "No results found."}, nil } var sb strings.Builder sb.WriteString(fmt.Sprintf("Search results for %q:\n\n", query)) for i, r := range results { sb.WriteString(fmt.Sprintf("%d. %s\n %s\n %s\n\n", i+1, r.Title, r.URL, r.Snippet)) } return map[string]interface{}{"content": strings.TrimSpace(sb.String())}, nil } // ── Web Fetch (quick) ────────────────────────────────────── func htmlToText(html string) string { for _, tag := range []string{"