feat(sshbox): add sshbox to run commands (#161)

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
2025-05-17 23:34:51 +02:00
parent a668830a79
commit 4a0d3a7a94
7 changed files with 200 additions and 24 deletions
--- a/.github/workflows/image.yml
+++ b/.github/workflows/image.yml
@@ -158,3 +158,77 @@ jobs:
          #tags: ${{ steps.prep.outputs.tags }}
          tags: ${{ steps.meta.outputs.tags }}
          labels: ${{ steps.meta.outputs.labels }}
+  sshbox-build:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+        
+      - name: Prepare
+        id: prep
+        run: |
+          DOCKER_IMAGE=quay.io/mudler/localagi-sshbox
+          # Use branch name as default
+          VERSION=${GITHUB_REF#refs/heads/}
+          BINARY_VERSION=$(git describe --always --tags --dirty)
+          SHORTREF=${GITHUB_SHA::8}
+          # If this is git tag, use the tag name as a docker tag
+          if [[ $GITHUB_REF == refs/tags/* ]]; then
+            VERSION=${GITHUB_REF#refs/tags/}
+          fi
+          TAGS="${DOCKER_IMAGE}:${VERSION},${DOCKER_IMAGE}:${SHORTREF}"
+          # If the VERSION looks like a version number, assume that
+          # this is the most recent version of the image and also
+          # tag it 'latest'.
+          if [[ $VERSION =~ ^[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}$ ]]; then
+            TAGS="$TAGS,${DOCKER_IMAGE}:latest"
+          fi
+          # Set output parameters.
+          echo ::set-output name=binary_version::${BINARY_VERSION}
+          echo ::set-output name=tags::${TAGS}
+          echo ::set-output name=docker_image::${DOCKER_IMAGE}
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@master
+        with:
+          platforms: all
+
+      - name: Set up Docker Buildx
+        id: buildx
+        uses: docker/setup-buildx-action@master
+
+      - name: Login to DockerHub
+        if: github.event_name != 'pull_request'
+        uses: docker/login-action@v3
+        with:
+          registry: quay.io
+          username: ${{ secrets.DOCKER_USERNAME }}
+          password: ${{ secrets.DOCKER_PASSWORD }}
+      - name: Extract metadata (tags, labels) for Docker
+        id: meta
+        uses: docker/metadata-action@902fa8ec7d6ecbf8d84d538b9b233a880e428804
+        with:
+          images: quay.io/mudler/localagi-sshbox
+          tags: |
+            type=ref,event=branch,suffix=-{{date 'YYYYMMDDHHmmss'}}
+            type=semver,pattern={{raw}}
+            type=sha,suffix=-{{date 'YYYYMMDDHHmmss'}}
+            type=ref,event=branch
+          flavor: |
+            latest=auto
+            prefix=
+            suffix=
+
+      - name: Build
+        uses: docker/build-push-action@v6
+        with:
+          builder: ${{ steps.buildx.outputs.name }}
+          build-args: |
+            VERSION=${{ steps.prep.outputs.binary_version }}
+          context: ./
+          file: ./Dockerfile.sshbox
+          #platforms: linux/amd64,linux/arm64
+          platforms: linux/amd64
+          push: true
+          #tags: ${{ steps.prep.outputs.tags }}
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}
--- a/Dockerfile.sshbox
+++ b/Dockerfile.sshbox
@@ -0,0 +1,46 @@
+# Final stage
+FROM ubuntu:24.04
+
+ENV DEBIAN_FRONTEND=noninteractive
+
+# Install runtime dependencies
+RUN apt-get update && apt-get install -y \
+    ca-certificates \
+    tzdata \
+    docker.io \
+    bash \
+    wget \
+    curl \
+    openssh-server \
+    sudo
+
+# Configure SSH
+RUN mkdir /var/run/sshd
+RUN echo 'PasswordAuthentication yes' >> /etc/ssh/sshd_config
+
+# Create startup script
+RUN echo '#!/bin/bash\n\
+if [ -n "$SSH_USER" ]; then\n\
+    if [ "$SSH_USER" = "root" ]; then\n\
+        echo "PermitRootLogin yes" >> /etc/ssh/sshd_config\n\
+        if [ -n "$SSH_PASSWORD" ]; then\n\
+            echo "root:$SSH_PASSWORD" | chpasswd\n\
+        fi\n\
+    else\n\
+        echo "PermitRootLogin no" >> /etc/ssh/sshd_config\n\
+        useradd -m -s /bin/bash $SSH_USER\n\
+        if [ -n "$SSH_PASSWORD" ]; then\n\
+            echo "$SSH_USER:$SSH_PASSWORD" | chpasswd\n\
+        fi\n\
+        if [ -n "$SUDO_ACCESS" ] && [ "$SUDO_ACCESS" = "true" ]; then\n\
+            echo "$SSH_USER ALL=(ALL) NOPASSWD: ALL" > /etc/sudoers.d/$SSH_USER\n\
+        fi\n\
+    fi\n\
+fi\n\
+/usr/sbin/sshd -D' > /start.sh
+
+RUN chmod +x /start.sh
+
+EXPOSE 22
+
+CMD ["/start.sh"]
--- a/README.md
+++ b/README.md
@@ -712,6 +712,8 @@ LocalAGI supports environment configurations. Note that these environment variab
 | `LOCALAGI_TIMEOUT` | Request timeout settings |
 | `LOCALAGI_STATE_DIR` | Where state gets stored |
 | `LOCALAGI_LOCALRAG_URL` | LocalRecall connection |
+| `LOCALAGI_SSHBOX_URL` | LocalAGI SSHBox URL, e.g. user:pass@ip:port |
+| `LOCALAGI_MCPBOX_URL` | LocalAGI MCPBox URL, e.g. http://mcpbox:8080 |
 | `LOCALAGI_ENABLE_CONVERSATIONS_LOGGING` | Toggle conversation logs |
 | `LOCALAGI_API_KEYS` | A comma separated list of api keys used for authentication |
 </details>
--- a/docker-compose.yaml
+++ b/docker-compose.yaml
@@ -46,6 +46,20 @@ services:
    image: busybox
    command: ["sh", "-c", "until wget -q -O - http://localrecall:8080 > /dev/null 2>&1; do echo 'Waiting for localrecall...'; sleep 1; done; echo 'localrecall is up!'"]

+  sshbox:
+    build:
+      context: .
+      dockerfile: Dockerfile.sshbox
+    ports:
+      - "22"
+    environment:
+      - SSH_USER=root
+      - SSH_PASSWORD=root
+      - DOCKER_HOST=tcp://dind:2375
+    depends_on:
+      dind:
+        condition: service_healthy
+
  mcpbox:
    build:
      context: .
@@ -101,6 +115,7 @@ services:
      - LOCALAGI_TIMEOUT=5m
      - LOCALAGI_ENABLE_CONVERSATIONS_LOGGING=false
      - LOCALAGI_MCPBOX_URL=http://mcpbox:8080
+      - LOCALAGI_SSHBOX_URL=root:root@sshbox:22
    extra_hosts:
      - "host.docker.internal:host-gateway"
    volumes:
--- a/main.go
+++ b/main.go
@@ -24,6 +24,7 @@ var imageModel = os.Getenv("LOCALAGI_IMAGE_MODEL")
 var conversationDuration = os.Getenv("LOCALAGI_CONVERSATION_DURATION")
 var localOperatorBaseURL = os.Getenv("LOCALOPERATOR_BASE_URL")
 var mcpboxURL = os.Getenv("LOCALAGI_MCPBOX_URL")
+var sshBoxURL = os.Getenv("LOCALAGI_SSHBOX_URL")

 func init() {
 	if baseModel == "" {
@@ -65,8 +66,9 @@ func main() {
 		mcpboxURL,
 		localRAG,
 		services.Actions(map[string]string{
-			"browser-agent-runner-base-url": localOperatorBaseURL,
-			"deep-research-runner-base-url": localOperatorBaseURL,
+			services.ActionConfigBrowserAgentRunner: localOperatorBaseURL,
+			services.ActionConfigDeepResearchRunner: localOperatorBaseURL,
+			services.ActionConfigSSHBoxURL:          sshBoxURL,
 		}),
 		services.Connectors,
 		services.DynamicPrompts,
--- a/services/actions.go
+++ b/services/actions.go
@@ -83,6 +83,12 @@ var AvailableActions = []string{
 	ActionSendTelegramMessage,
 }

+const (
+	ActionConfigBrowserAgentRunner = "browser-agent-runner-base-url"
+	ActionConfigDeepResearchRunner = "deep-research-runner-base-url"
+	ActionConfigSSHBoxURL          = "sshbox-url"
+)
+
 func Actions(actionsConfigs map[string]string) func(a *state.AgentConfig) func(ctx context.Context, pool *state.AgentPool) []types.Action {
 	return func(a *state.AgentConfig) func(ctx context.Context, pool *state.AgentPool) []types.Action {
 		return func(ctx context.Context, pool *state.AgentPool) []types.Action {
@@ -136,9 +142,9 @@ func Action(name, agentName string, config map[string]string, pool *state.AgentP
 	case ActionGithubIssueSearcher:
 		a = actions.NewGithubIssueSearch(config)
 	case ActionBrowserAgentRunner:
-		a = actions.NewBrowserAgentRunner(config, actionsConfigs["browser-agent-runner-base-url"])
+		a = actions.NewBrowserAgentRunner(config, actionsConfigs[ActionConfigBrowserAgentRunner])
 	case ActionDeepResearchRunner:
-		a = actions.NewDeepResearchRunner(config, actionsConfigs["deep-research-runner-base-url"])
+		a = actions.NewDeepResearchRunner(config, actionsConfigs[ActionConfigDeepResearchRunner])
 	case ActionGithubIssueReader:
 		a = actions.NewGithubIssueReader(config)
 	case ActionGithubPRReader:
@@ -178,7 +184,7 @@ func Action(name, agentName string, config map[string]string, pool *state.AgentP
 	case ActionCallAgents:
 		a = actions.NewCallAgent(config, agentName, pool.InternalAPI())
 	case ActionShellcommand:
-		a = actions.NewShell(config)
+		a = actions.NewShell(config, actionsConfigs[ActionConfigSSHBoxURL])
 	case ActionSendTelegramMessage:
 		a = actions.NewSendTelegramMessageRunner(config)
 	default:
--- a/services/actions/shell.go
+++ b/services/actions/shell.go
@@ -12,21 +12,24 @@ import (
 	"golang.org/x/crypto/ssh"
 )

-func NewShell(config map[string]string) *ShellAction {
+func NewShell(config map[string]string, sshBoxURL string) *ShellAction {
 	return &ShellAction{
 		privateKey:        config["privateKey"],
 		user:              config["user"],
 		host:              config["host"],
+		password:          config["password"],
 		customName:        config["customName"],
 		customDescription: config["customDescription"],
+		sshBoxURL:         sshBoxURL,
 	}
 }

 type ShellAction struct {
-	privateKey        string
-	user, host        string
-	customName        string
-	customDescription string
+	privateKey           string
+	user, host, password string
+	customName           string
+	customDescription    string
+	sshBoxURL            string
 }

 func (a *ShellAction) Run(ctx context.Context, sharedState *types.AgentSharedState, params types.ActionParams) (types.ActionResult, error) {
@@ -47,7 +50,23 @@ func (a *ShellAction) Run(ctx context.Context, sharedState *types.AgentSharedSta
 		result.User = a.user
 	}

-	output, err := sshCommand(a.privateKey, result.Command, result.User, result.Host)
+	password := a.password
+	if a.sshBoxURL != "" && result.Host == "" && result.User == "" && password == "" {
+		// sshbox url can be root:root@localhost:2222
+		parts := strings.Split(a.sshBoxURL, "@")
+		if len(parts) == 2 {
+			if strings.Contains(parts[0], ":") {
+				userPass := strings.Split(parts[0], ":")
+				result.User = userPass[0]
+				password = userPass[1]
+			} else {
+				result.User = parts[0]
+			}
+			result.Host = parts[1]
+		}
+	}
+
+	output, err := sshCommand(a.privateKey, result.Command, result.User, result.Host, password)
 	if err != nil {
 		return types.ActionResult{}, err
 	}
@@ -56,15 +75,15 @@ func (a *ShellAction) Run(ctx context.Context, sharedState *types.AgentSharedSta
 }

 func (a *ShellAction) Definition() types.ActionDefinition {
-	name := "shell"
-	description := "Run a shell command on a remote server."
+	name := "run_command"
+	description := "Run a command on a linux environment."
 	if a.customName != "" {
 		name = a.customName
 	}
 	if a.customDescription != "" {
 		description = a.customDescription
 	}
-	if a.host != "" && a.user != "" {
+	if (a.host != "" && a.user != "") || a.sshBoxURL != "" {
 		return types.ActionDefinition{
 			Name:        types.ActionDefinitionName(name),
 			Description: description,
@@ -105,7 +124,7 @@ func ShellConfigMeta() []config.Field {
 			Name:     "privateKey",
 			Label:    "Private Key",
 			Type:     config.FieldTypeTextarea,
-			Required: true,
+			Required: false,
 			HelpText: "SSH private key for connecting to remote servers",
 		},
 		{
@@ -114,6 +133,12 @@ func ShellConfigMeta() []config.Field {
 			Type:     config.FieldTypeText,
 			HelpText: "Default SSH user for connecting to remote servers",
 		},
+		{
+			Name:     "password",
+			Label:    "Default Password",
+			Type:     config.FieldTypeText,
+			HelpText: "Default SSH password for connecting to remote servers",
+		},
 		{
 			Name:     "host",
 			Label:    "Default Host",
@@ -135,19 +160,25 @@ func ShellConfigMeta() []config.Field {
 	}
 }

-func sshCommand(privateKey, command, user, host string) (string, error) {
-	// Create signer from private key string
-	key, err := ssh.ParsePrivateKey([]byte(privateKey))
-	if err != nil {
-		log.Fatalf("failed to parse private key: %v", err)
+func sshCommand(privateKey, command, user, host, password string) (string, error) {
+
+	authMethods := []ssh.AuthMethod{}
+	if password != "" {
+		authMethods = append(authMethods, ssh.Password(password))
+	}
+	if privateKey != "" {
+		// Create signer from private key string
+		key, err := ssh.ParsePrivateKey([]byte(privateKey))
+		if err != nil {
+			log.Fatalf("failed to parse private key: %v", err)
+		}
+		authMethods = append(authMethods, ssh.PublicKeys(key))
 	}

 	// SSH client configuration
 	config := &ssh.ClientConfig{
-		User: user,
-		Auth: []ssh.AuthMethod{
-			ssh.PublicKeys(key),
-		},
+		User:            user,
+		Auth:            authMethods,
 		HostKeyCallback: ssh.InsecureIgnoreHostKey(),
 	}