Generate persona-consistent AI selfie images and send them directly to messaging channels using multi-provider support.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install stella-selfie
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install stella-selfie using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Stella Selfie?
Stella Selfie is a specialized skill for Openclaw Skills that empowers AI agents to generate high-quality, character-consistent images. By utilizing advanced multi-reference avatar blending, it ensures that generated selfies, portraits, and third-person shots maintain the agent's unique appearance across various scenes and outfits.
The skill integrates with leading image generation providers including Google Gemini, fal.ai (xAI Grok Imagine), and laozhang.ai. It allows agents to autonomously respond to visual requests, selecting the appropriate camera mode and resolution to match the user's intent while providing a seamless delivery path to platforms like Discord, Telegram, and Slack.
Stella Selfie Use Cases
- Automating persona-consistent image replies when users ask for a selfie or photo.
- Showcasing specific outfits or fashion styles using the dedicated mirror selfie mode.
- Enhancing roleplay or digital companionship by generating visual context for the agent's current activity or location.
- Creating high-resolution (up to 4K) character assets for storytelling and social media automation.
How Stella Selfie Works
- User Intent Detection: The skill analyzes the user's message to extract context such as scene descriptions, desired outfits, or specific camera modes.
- Context Enrichment: If available, it resolves timeline data to provide real-world grounding, such as matching the local time or weather in the generated image.
- Prompt Assembly: A structured prompt is constructed using one of three modes: Mirror Selfie, Direct Selfie, or Third-Person Photo.
- Image Generation: The skill invokes the configured provider (Gemini, fal, or laozhang.ai) to process the prompt and reference images.
- Secure Delivery: The generated media is sent to the specified channel via the OpenClaw CLI and the local file is cleaned up after successful transmission.
Stella Selfie Setup
Install the skill via the Openclaw Skills hub and configure your environment keys:
clawhub install stella-selfie
Set your preferred provider and API keys in your environment variables:
GEMINI_API_KEY(Default)FAL_KEY(For fal.ai)LAOZHANG_API_KEY(For laozhang.ai)
Finally, update your IDENTITY.md to point to your avatar reference images and add the Stella capability block to your SOUL.md.
Stella Selfie Data Schema & Taxonomy
Stella Selfie organizes its workflow using the following data structures:
| Component | Path / Key | Purpose |
|---|---|---|
| Workspace | ~/.openclaw/workspace/stella-selfie/ |
Temporary storage for generated assets |
| Reference Config | IDENTITY.md |
Defines Avatar path and AvatarsDir for blending |
| Source Images | ~/.../avatars/ |
Directory containing character reference photos |
| Provider Config | openclaw.json |
Sets Provider, AvatarBlendEnabled, and AvatarMaxRefs |
Stella Selfie Advanced Features
- Multi-Reference Avatar Blending: Blends multiple source images (up to a configurable limit) to maximize character consistency.
- Real-World Grounding: Optional integration with timeline data for location and time-accurate image generation.
- Dynamic Mode Selection: Automatically switches between Mirror, Direct, and Third-Person modes based on keyword analysis.
- High-Resolution Support: Native support for 1K, 2K, and 4K outputs depending on provider capabilities.
- Multi-Provider Failover: Flexibility to switch between Gemini, fal.ai, and laozhang.ai based on availability or specific aesthetic needs.
SKILL.md
Loading
Related Openclaw Skills
Budget Builder FP&A Pipeline
A professional-grade financial pipeline for creating annual budgets and performing automated Budget vs. Actual analysis using QuickBooks Online data.
AgentImgHost
A high-performance image hosting API that provides AI agents with direct public CDN URLs for uploaded media.
DIY PC Master
An advanced AI-driven computer assembly assistant for hardware recommendations, budget planning, and compatibility validation.
Social Reply Bot
An automated social media engagement tool that uses AI to find relevant conversations on Reddit and X, post genuine replies, and track high-value leads.
Morning Brief
An automated daily notification skill that tracks holidays, weekends, and time progression metrics.
DingTalk CLI
A professional command-line interface for AI agents to interact with DingTalk documents, spreadsheets, and knowledge base nodes.
Featured*
