Merge pull request #622 from docker/add-skills

Add skills command for AI coding assistants

Author: ericcurtin

cmd/cli/commands/root.go

@@ -94,6 +94,7 @@ func NewRootCmd(cli *command.DockerCli) *cobra.Command {
 		newRestartRunner(),
 		newReinstallRunner(),
 		newSearchCmd(),
+		newSkillsCmd(),
 	)
 
 	// Commands that require a running model runner. These are wrapped to ensure the standalone runner is available.

cmd/cli/commands/skills.go

@@ -0,0 +1,145 @@
+package commands
+
+import (
+	"embed"
+	"fmt"
+	"io/fs"
+	"os"
+	"path/filepath"
+
+	"github.com/docker/model-runner/cmd/cli/commands/completion"
+	"github.com/spf13/cobra"
+)
+
+//go:embed skills/*
+var skillsFS embed.FS
+
+type skillsOptions struct {
+	codex    bool
+	claude   bool
+	opencode bool
+	dest     string
+	force    bool
+}
+
+func newSkillsCmd() *cobra.Command {
+	opts := &skillsOptions{}
+
+	c := &cobra.Command{
+		Use:   "skills",
+		Short: "Install Docker Model Runner skills for AI coding assistants",
+		Long: `Install Docker Model Runner skills for AI coding assistants.
+
+Skills are configuration files that help AI coding assistants understand
+how to use Docker Model Runner effectively for local model inference.
+
+Supported targets:
+  --codex     Install to ~/.codex/skills (OpenAI Codex CLI)
+  --claude    Install to ~/.claude/skills (Claude Code)
+  --opencode  Install to ~/.config/opencode/skills (OpenCode)
+  --dest      Install to a custom directory
+
+Example:
+  docker model skills --claude
+  docker model skills --codex --claude
+  docker model skills --dest /path/to/skills`,
+		RunE: func(cmd *cobra.Command, args []string) error {
+			return runSkills(cmd, opts)
+		},
+		ValidArgsFunction: completion.NoComplete,
+	}
+
+	c.Flags().BoolVar(&opts.codex, "codex", false, "Install skills for OpenAI Codex CLI (~/.codex/skills)")
+	c.Flags().BoolVar(&opts.claude, "claude", false, "Install skills for Claude Code (~/.claude/skills)")
+	c.Flags().BoolVar(&opts.opencode, "opencode", false, "Install skills for OpenCode (~/.config/opencode/skills)")
+	c.Flags().StringVar(&opts.dest, "dest", "", "Install skills to a custom directory")
+	c.Flags().BoolVarP(&opts.force, "force", "f", false, "Overwrite existing skills without prompting")
+
+	return c
+}
+
+func runSkills(cmd *cobra.Command, opts *skillsOptions) error {
+	// Collect target directories
+	var targets []string
+	homeDir, err := os.UserHomeDir()
+	if err != nil {
+		return fmt.Errorf("failed to get home directory: %w", err)
+	}
+
+	if opts.codex {
+		targets = append(targets, filepath.Join(homeDir, ".codex", "skills"))
+	}
+	if opts.claude {
+		targets = append(targets, filepath.Join(homeDir, ".claude", "skills"))
+	}
+	if opts.opencode {
+		targets = append(targets, filepath.Join(homeDir, ".config", "opencode", "skills"))
+	}
+	if opts.dest != "" {
+		targets = append(targets, opts.dest)
+	}
+
+	if len(targets) == 0 {
+		return fmt.Errorf("no target specified. Use --codex, --claude, --opencode, or --dest")
+	}
+
+	// Install skills to each target
+	for _, target := range targets {
+		if err := installSkills(cmd, target, opts.force); err != nil {
+			return fmt.Errorf("failed to install skills to %s: %w", target, err)
+		}
+		cmd.Printf("Installed Docker Model Runner skills to %s\n", target)
+	}
+
+	return nil
+}
+
+func installSkills(cmd *cobra.Command, targetDir string, force bool) error {
+	// Walk through embedded skills directory
+	return fs.WalkDir(skillsFS, "skills", func(path string, d fs.DirEntry, err error) error {
+		if err != nil {
+			return err
+		}
+
+		// Skip the root "skills" directory itself
+		if path == "skills" {
+			return nil
+		}
+
+		// Calculate the relative path from "skills/"
+		relPath, err := filepath.Rel("skills", path)
+		if err != nil {
+			return err
+		}
+
+		destPath := filepath.Join(targetDir, relPath)
+
+		if d.IsDir() {
+			// Create directory
+			return os.MkdirAll(destPath, 0755)
+		}
+
+		// Check if file exists and handle force flag
+		if _, err := os.Stat(destPath); err == nil && !force {
+			return fmt.Errorf("file already exists: %s (use --force to overwrite)", destPath)
+		}
+
+		// Read the embedded file
+		content, err := skillsFS.ReadFile(path)
+		if err != nil {
+			return fmt.Errorf("failed to read embedded file %s: %w", path, err)
+		}
+
+		// Ensure parent directory exists
+		if err := os.MkdirAll(filepath.Dir(destPath), 0755); err != nil {
+			return fmt.Errorf("failed to create directory for %s: %w", destPath, err)
+		}
+
+		// Write the file
+		if err := os.WriteFile(destPath, content, 0644); err != nil {
+			return fmt.Errorf("failed to write file %s: %w", destPath, err)
+		}
+
+		return nil
+	})
+}

cmd/cli/commands/skills/docker-model-runner/SKILL.md

@@ -0,0 +1,87 @@
+---
+name: docker-model-runner
+description: Skills for using Docker Model Runner to run local LLM inference
+---
+
+# Docker Model Runner
+
+Docker Model Runner (DMR) makes it easy to run AI models locally using Docker. This skill helps you effectively use Docker Model Runner for local LLM inference in your development workflow.
+
+## Workflow
+
+When helping users with local LLM inference using Docker Model Runner:
+
+1. **Check if Docker Model Runner is available** by running `docker model version`
+
+2. **List available models** with `docker model list` to see what's already pulled
+
+3. **Search for models** on Docker Hub or HuggingFace:
+   - `docker model search <query>` to find models
+   - Popular models include: `ai/gemma3`, `ai/llama3.2`, `ai/smollm2`, `ai/qwen3`
+
+4. **Pull models** before running: `docker model pull <model>`
+
+5. **Run models** for inference:
+   - One-time prompt: `docker model run ai/smollm2 "Your prompt here"`
+   - Interactive chat: `docker model run ai/smollm2`
+   - Pre-load model: `docker model run --detach ai/smollm2`
+
+6. **Use the OpenAI-compatible API** for programmatic access:
+   - Endpoint: `http://localhost:12434/engines/llama.cpp/v1/chat/completions`
+   - This is compatible with OpenAI client libraries
+
+## API Usage
+
+Docker Model Runner exposes an OpenAI-compatible REST API:
+
+```bash
+# Chat completions
+curl http://localhost:12434/engines/llama.cpp/v1/chat/completions \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "ai/smollm2",
+    "messages": [
+      {"role": "system", "content": "You are a helpful assistant."},
+      {"role": "user", "content": "Hello!"}
+    ]
+  }'
+```
+
+For Python with the OpenAI library:
+
+```python
+from openai import OpenAI
+
+client = OpenAI(
+    base_url="http://localhost:12434/engines/llama.cpp/v1",
+    api_key="not-needed"  # API key not required for local inference
+)
+
+response = client.chat.completions.create(
+    model="ai/smollm2",
+    messages=[{"role": "user", "content": "Hello!"}]
+)
+```
+
+## Key Commands
+
+| Command | Description |
+|---------|-------------|
+| `docker model run <model> [prompt]` | Run a model with optional prompt |
+| `docker model pull <model>` | Pull a model from registry |
+| `docker model list` | List downloaded models |
+| `docker model search <query>` | Search for models |
+| `docker model ps` | Show running models |
+| `docker model rm <model>` | Remove a model |
+| `docker model inspect <model>` | Show model details |
+
+## Best Practices
+
+- Use smaller models (like `ai/smollm2`) for faster responses during development
+- Pre-load models with `--detach` for better performance in scripts
+- Models stay loaded until another model is requested or timeout (5 min)
+- Use the OpenAI-compatible API for integration with existing tools
+
+## References
+
+See [references/docker-model-guide.md](references/docker-model-guide.md) for detailed documentation.