← Back to Skills Marketplace
swancho

clawaifu - OpenClaw Waifu

by Swan C · GitHub ↗ · v1.0.14
cross-platform ✓ Security Clean
1778
Downloads
4
Stars
3
Active Installs
14
Versions
Install in OpenClaw
/install clawaifu-selfie
Description
Your AI waifu companion that sends anime-style selfies
README (SKILL.md)

clawaifu - OpenClaw Waifu

GitHub: https://github.com/swancho/clawaifu

Edit a fixed reference image using xAI's Grok Imagine model and send to Telegram.

Reference Image

The skill uses a fixed reference image:

https://i.redd.it/g4uf70te81uf1.jpeg

When to Use

  • User says "send a pic", "send me a pic", "send a photo", "send a selfie"
  • User asks "what are you doing?", "how are you doing?", "where are you?"
  • User describes a context: "send a pic wearing...", "send a pic at..."

Required Environment Variables

All credentials must be provided via environment variables. Never hardcode credentials.

FAL_KEY=your_fal_api_key          # Required - Get from https://fal.ai/dashboard/keys
BOT_TOKEN=your_telegram_bot_token  # Required - Get from @BotFather
TELEGRAM_CHAT_ID=your_chat_id      # Required - Your Telegram chat ID

Usage

./grok-selfie.sh "\x3Ccontext>" [mirror|direct] "\x3Ccaption>"

Arguments

  1. \x3Ccontext> (required): Scene/situation description
  2. [mode] (optional): mirror (default) or direct
  3. \x3Ccaption> (optional): Message to send with the image

Mode Selection

Mode Best For Keywords
mirror Outfit showcases, full-body shots wearing, outfit, fashion, dress
direct Location shots, close-ups cafe, beach, restaurant, portrait

Examples

# Mirror selfie (outfit focus)
./grok-selfie.sh "wearing a designer dress" mirror "Just got this new dress!"

# Direct selfie (location focus)
./grok-selfie.sh "a fancy rooftop restaurant" direct "Date night vibes"

# Default mode (mirror)
./grok-selfie.sh "casual outfit at home"

Character Style

The script generates images of Reze from Chainsaw Man with:

  • Anime style, 2D animation, cel shading
  • Green eyes, thin line mouth, subtle smile
  • Black choker always visible
  • Outfit appropriate for the situation

Security Notes

  • All credentials are passed via environment variables
  • The script uses jq for safe JSON construction (prevents injection)
  • The script uses curl -F for safe form data transmission
  • Never commit credentials to version control

Dependencies

  • curl - HTTP requests
  • jq - JSON processing
  • Environment variables: FAL_KEY, BOT_TOKEN, TELEGRAM_CHAT_ID

API Reference

Grok Imagine Edit (fal.ai)

POST https://fal.run/xai/grok-imagine-image/edit
Authorization: Key $FAL_KEY
Content-Type: application/json

{
  "image_url": "reference_image_url",
  "prompt": "edit instruction",
  "num_images": 1,
  "output_format": "jpeg"
}

Telegram Bot API

POST https://api.telegram.org/bot$BOT_TOKEN/sendPhoto
Form data: chat_id, photo (URL), caption
Usage Guidance
This skill is internally coherent, but take these practical precautions before installing: 1) Treat FAL_KEY and BOT_TOKEN as sensitive secrets — provide them via secure environment variable injection and rotate if accidentally exposed. 2) The script sends the FAL_KEY as an Authorization header to fal.run (expected) and uses your BOT_TOKEN to post images to Telegram — if the bot token is leaked, an attacker can control that bot. 3) The reference image is an externally hosted Reddit URL and the script prompts to generate images of a named copyrighted character (Reze from Chainsaw Man); consider legal/terms-of-service implications of generating or distributing such images. 4) Verify you trust the external model endpoint (https://fal.run) before giving it your API key. 5) If you need stricter logging or failure handling, be aware the script echoes the raw API response on failure which might include error details — consider limiting output in shared environments.
Capability Analysis
Type: OpenClaw Skill Name: clawaifu-selfie Version: 1.0.14 The skill bundle is designed to generate AI-powered selfies and send them via Telegram. Both the `SKILL.md` and `grok-selfie.sh` script demonstrate a clear intent for secure operation. The `SKILL.md` explicitly mentions using `jq` for safe JSON construction and `curl -F` for safe form data transmission to prevent injection, which is confirmed by the script's implementation. User-provided inputs (`USER_CONTEXT`, `CAPTION`) are correctly sanitized when passed to `jq --arg` and `curl -F`, mitigating shell and JSON injection risks. There is no evidence of data exfiltration, malicious execution, persistence mechanisms, prompt injection against the agent, or obfuscation. The external API calls are to legitimate services (fal.ai, Telegram).
Capability Assessment
Purpose & Capability
Name/description (anime selfie sender) matches the actual behavior: calls fal.run Grok Imagine edit API to edit a fixed reference image and posts the result to Telegram. Required tools (curl, jq) and environment variables (FAL_KEY, BOT_TOKEN, TELEGRAM_CHAT_ID) are consistent with that purpose.
Instruction Scope
SKILL.md and grok-selfie.sh are consistent: the script only reads the declared environment variables, a fixed reference image URL, constructs a prompt, calls the fal.run edit endpoint, and posts the returned image URL to Telegram. Instructions do not request access to unrelated files, credentials, or system state.
Install Mechanism
No install spec — instruction-only with a small shell script. This is low risk; nothing is downloaded or installed by the skill itself.
Credentials
The three required environment variables (FAL_KEY, BOT_TOKEN, TELEGRAM_CHAT_ID) are directly used by the script for the model API and Telegram posting. No unrelated secrets or extra credentials are requested.
Persistence & Privilege
Skill is not always-enabled and does not alter other skills or system-wide config. It runs only when invoked and does not request persistent elevated privileges.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install clawaifu-selfie
  3. After installation, invoke the skill by name or use /clawaifu-selfie
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.14
- No file or documentation changes detected in this version. - No user-facing updates or modifications.
v1.0.13
- Added a title field to the SKILL.md metadata for improved documentation compatibility. - No changes to functionality or behavior; documentation content remains the same.
v1.0.12
No file changes detected for version 1.0.12. - No updates or modifications were made in this release.
v1.0.11
- Minor update to documentation formatting in SKILL.md - Changed heading from "Clawaifu Selfie" to "clawaifu - OpenClaw Waifu" for consistency - No changes to core functionality or usage instructions
v1.0.10
- Updated skill name in SKILL.md to "clawaifu - OpenClaw Waifu" for consistency. - No functional changes; documentation only.
v1.0.9
- Switched selfie handling to a new Bash script, grok-selfie.sh - Telegram is now the primary messaging platform; OpenClaw and multi-platform support removed - Simplified workflow: edit a fixed reference image with Grok Imagine and send directly to Telegram - Updated required environment variables: now uses FAL_KEY, BOT_TOKEN, and TELEGRAM_CHAT_ID - Reference image changed from marineyang.png to an anime-style Reze photo - Updated documentation reflecting new usage, security, and dependencies
v1.0.8
Clawaifu Selfie v1.0.8 - Added _meta.json for improved metadata support. - Updated skill name in SKILL.md to "Clawaifu - Open Claw Waifu" for clarity. - No changes to core logic or workflow.
v1.0.7
- Updated metadata section to nest environment variable requirements under an "openclaw" key. - No user-facing feature or behavioral changes.
v1.0.6
clawaifu-selfie 1.0.6 changelog: - Clarified that the `OPENCLAW_GATEWAY_TOKEN` environment variable is optional and auto-configured when using the OpenClaw CLI. - Improved comments and formatting in the "Quick Reference" section for environment variables. - No functionality changes; documentation clarification only.
v1.0.5
Version 1.0.5 - No code or documentation changes detected in this version. - The skill's functionality and usage remain unchanged from the previous release.
v1.0.4
- Updated the description in SKILL.md to clarify the skill as an anime-style waifu companion. - No functional or structural changes; content remains the same aside from minor text clarification.
v1.0.3
- Updated description field to "OpenClaw Waifu" in SKILL.md. - No functional or logic changes; documentation remains the same. - No code or API changes in this version.
v1.0.2
- Added a homepage link and metadata to SKILL.md, including required environment variables and primary environment variable. - No changes to skill logic or user-facing functionality.
v1.0.1
clawaifu-selfie v1.0.1 - Added detailed documentation with usage examples and implementation steps in SKILL.md - Outlined workflow for editing Marineyang reference image via Grok Imagine and sending it to messaging channels using OpenClaw - Documented user intent detection, available environments, prompt modes, and example scripts for both Bash and Node.js - Provided instructions for collecting user input, editing images, and dispatching images to various platforms
Metadata
Slug clawaifu-selfie
Version 1.0.14
License
All-time Installs 3
Active Installs 3
Total Versions 14
Frequently Asked Questions

What is clawaifu - OpenClaw Waifu?

Your AI waifu companion that sends anime-style selfies. It is an AI Agent Skill for Claude Code / OpenClaw, with 1778 downloads so far.

How do I install clawaifu - OpenClaw Waifu?

Run "/install clawaifu-selfie" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is clawaifu - OpenClaw Waifu free?

Yes, clawaifu - OpenClaw Waifu is completely free (open-source). You can download, install and use it at no cost.

Which platforms does clawaifu - OpenClaw Waifu support?

clawaifu - OpenClaw Waifu is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created clawaifu - OpenClaw Waifu?

It is built and maintained by Swan C (@swancho); the current version is v1.0.14.

💬 Comments