← 返回 Skills 市场

Fitness Recipes AI

Name: Fitness Recipes AI
Author: g0atfac3

作者 G0atfac3 · GitHub ↗ · v1.0.0 · MIT-0

cross-platform ⚠ suspicious

274

总下载

当前安装

版本数

在 OpenClaw 中安装

/install g0atbot-fitness-recipes

功能描述

AI tool that creates viral fitness recipe videos with AI-generated images, voiceovers, renders, and optional TikTok auto-posting.

使用说明 (SKILL.md)

Fitness Recipes AI

AI-powered TikTok video generator for fitness content. Creates viral recipe videos from a database, generates AI images via fal.ai, adds voiceover with ElevenLabs, renders video with Shotstack, and posts to TikTok.

What It Does

Maintains a recipe database (or fetches from source)
Generates AI food images using fal.ai
Creates voiceover scripts and generates audio with ElevenLabs
Renders complete videos using Shotstack API
Posts to TikTok via Postiz (optional) or exports for manual upload

Setup

Prerequisites

fal.ai API Key - Get at https://fal.ai
ElevenLabs API Key - Get at https://elevenlabs.io
Shotstack API Key - Get at https://shotstack.io
Postiz API Key (optional) - For auto-posting to TikTok

Environment Variables

export FAL_API_KEY="your_fal_key"
export ELEVENLABS_API_KEY="your_elevenlabs_key"
export SHOTSTACK_API_KEY="your_shotstack_key"
export POSTIZ_API_KEY="your_postiz_key"  # optional

Installation

pip install requests

Usage

Generate a single video

python generate_videos.py --recipe "Chicken Breast Recipe" --output my_video.mp4

Run daily batch

python daily_batch.py --count 10

Custom pipeline

from run_pipeline import Pipeline

pipeline = Pipeline()
video_path = pipeline.run(recipe_name="Keto Salmon", style="fitness")

Configuration

Edit config.py or set environment variables:

DEFAULT_VOICE - ElevenLabs voice ID
VIDEO_QUALITY - 1080p or 720p
OUTPUT_DIR - Where videos are saved

Cost Per Video

Component	Cost
fal.ai image	~$0.01
ElevenLabs voice	~$0.02
Shotstack render	~$0.10
Total	~$0.13

Example Output

Videos are saved to output/ directory with format: {recipe_name}_{timestamp}.mp4

Files

generate_videos.py - Main video generator
run_pipeline.py - Full pipeline orchestrator
fal_client.py - Image generation
elevenlabs_client.py - Voice generation
shotstack_client.py - Video rendering
daily_batch.py - Batch processor

安全使用建议

Do not run this skill as-is. Before installing or executing it: 1) Treat any hardcoded secret as compromised — rotate the exposed ElevenLabs key immediately if it's yours and remove secrets from code. 2) Request a corrected package that: removes hardcoded keys, documents and installs required system binaries (ffmpeg) and Python packages (Pillow, requests), and uses environment variables consistently. 3) Fix the syntax/implementation errors (malformed JSON/payload in fal_client.py and TODO stubs). 4) Remove or parameterize hardcoded absolute paths and avoid cron examples that run as root; let users choose install/output directories. 5) If you intend to auto-post to TikTok, verify the Postiz integration and limit token scope. If the publisher cannot address these issues or you can't audit the code yourself, avoid installing or running this skill and prefer a vetted alternative.

功能分析

Type: OpenClaw Skill Name: g0atbot-fitness-recipes Version: 1.0.0 The skill bundle contains a hardcoded ElevenLabs API key (sk_64a3...) in 'create_videos.py', which is a significant security risk and potential credential leak. Additionally, 'shotstack_client.py' attempts to pass local filesystem paths to a remote API using 'file://' URIs, which is a functional vulnerability (potential SSRF or path disclosure) if the remote service attempts to resolve them. While the code aligns with the stated purpose of video generation, these high-risk coding practices warrant a suspicious classification.

能力评估

⚠ Purpose & Capability

Name/description claim a video-generation pipeline using fal.ai, ElevenLabs, Shotstack and optional TikTok posting — the code does implement those components. However there are mismatches: several files call ffmpeg and Pillow (PIL) but SKILL.md only instructs `pip install requests`. _meta.json lists required env vars but create_videos.py embeds a hardcoded ElevenLabs API key and uses an absolute OUTPUT_DIR (/Users/...), while other files use Path.home()/clawd. The skill asks for no required binaries in registry metadata but relies on ffmpeg and cron examples, which is inconsistent.

⚠ Instruction Scope

SKILL.md instructs setting FAL/ELEVENLABS/SHOTSTACK env vars and a simple pip install of requests. The code, however, reads env vars in some modules but also hardcodes a secret in create_videos.py. The instructions don't document required system binaries (ffmpeg) or Python packages (Pillow). Some modules attempt network requests to external services and write to user's home directories and hardcoded paths; daily_batch.py suggests running as root in a cron entry. There are also TODOs and placeholder implementations in several files, and one module (fal_client.py) contains a malformed JSON payload line (syntax error), indicating the skill's runtime behavior is inconsistent with the documentation.

ℹ Install Mechanism

This is an instruction-only skill (no installer) which is lower risk from arbitrary downloads. However, the code writes files beneath the user's home dir and expects external system tools (ffmpeg) and packages (Pillow) that are not declared. Missing installation guidance for binaries is a usability and safety concern (scripts call ffmpeg via subprocess).

⚠ Credentials

SKILL.md and _meta.json declare FAL_API_KEY, ELEVENLABS_API_KEY, and SHOTSTACK_API_KEY (plus optional POSTIZ). Those env vars are appropriate for the described integration, but create_videos.py embeds an explicit ElevenLabs API key in source — a clear secret leak and inconsistency (some files use env vars, one file ignores them). The hardcoded key is disproportionate and dangerous: if it's valid, it exposes an account and allows unauthorized use; even if stale, embedding secrets in source is insecure. The number of env vars is reasonable for the purpose, but credential handling is sloppy.

✓ Persistence & Privilege

The skill does not request always:true, does not modify other skills, and is user-invocable. It does suggest adding a cron job (daily_batch) which is a user action; this is not an automatic privilege escalation by the skill itself.

如何使用

确保已安装 OpenClaw（本地或 Docker 部署）
在对话框中输入安装命令：/install g0atbot-fitness-recipes
安装完成后，直接呼叫该 Skill 的名称或使用 /g0atbot-fitness-recipes 触发
根据 Skill 的参数说明提供必要输入，即可获得结构化输出

版本历史

v1.0.0

Fitness Recipes AI 1.0.0 – Initial Release - Launches an AI-powered tool to create viral fitness recipe videos for TikTok. - Automates image generation (fal.ai), voiceover (ElevenLabs), video rendering (Shotstack), and TikTok posting (Postiz or manual export). - Supports batch and single-video creation workflows. - Includes configuration options for voice, quality, and output location. - Documents setup requirements, API keys, and estimated per-video costs.

元数据

Slug g0atbot-fitness-recipes

版本 1.0.0

许可证 MIT-0

累计安装 0

当前安装数 0

历史版本数 1

常见问题