← 返回 Skills 市场
320
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install bnbot-mascot
功能描述
Generate BNBOT lobster-bot mascot images in consistent style using reference images. Use when the user asks to generate mascot illustrations, character poses...
使用说明 (SKILL.md)
BNBOT Mascot Generator
Generate BNBOT's lobster-bot mascot in various poses and scenarios while maintaining consistent character design, using existing reference images for style matching.
How It Works
- Takes a reference image from the pre-approved set (V6/V7/V8 candidates)
- Sends it to Gemini as style reference along with the action prompt
- Gemini generates a new image matching the style
- Green-screen chromakey removes background for transparent PNG
Prerequisites
- Python:
/Users/jacklee/Projects/BNBOT/backend/.venv/bin/python - Dependencies:
google-genai,numpy,Pillow(already installed) - API key:
GOOGLE_AI_API_KEYin/Users/jacklee/Projects/BNBOT/.env - Reference images in
~/.claude/skills/bnbot-mascot/references/
Usage
Quick generation via CLI
/Users/jacklee/Projects/BNBOT/backend/.venv/bin/python \
~/.claude/skills/bnbot-mascot/scripts/generate_mascot.py \
"waving hello cheerfully" \
output.png \
full-body-front
In Python script
import sys
sys.path.insert(0, "/Users/jacklee/.claude/skills/bnbot-mascot/scripts")
from generate_mascot import generate_mascot
# Full body mascot
result = generate_mascot(
action="waving hello cheerfully",
output_path="mascot_wave.png",
reference="full-body-front", # style reference
background="transparent", # or "black", "white"
)
# Logo/icon version (head + claws only)
result = generate_mascot(
action="looking excited",
output_path="mascot_icon.png",
reference="logo-head-claws",
)
Available References
| Reference | Description | Best for |
|---|---|---|
full-body-front |
Front facing, symmetric, claws raised | General mascot, banners |
full-body-side |
Playful side pose, one claw up | Dynamic illustrations |
full-body-action |
Action pose, LED heart eye | Exciting/promo content |
full-body-cute |
Cute pose, CRT head, winking | Friendly/approachable content |
logo-head-claws |
Head + claws only, no body | Icons, logos, small sizes |
Action Ideas
For tweets / social media
waving hello cheerfully- 打招呼giving a thumbs up with one claw- 点赞holding a golden trophy- 庆祝成就sitting at a computer typing- 工作中wearing sunglasses, looking cool- 耍酷holding a megaphone, announcing something- 宣传celebrating with confetti- 庆祝thinking with one claw on chin- 思考
For stickers / reactions
laughing happily- 开心crying with pixel tears on screen- 伤心angry with pixel fire eyes- 生气sleeping with pixel ZZZ on screen- 睡觉shocked with pixel exclamation marks- 惊讶heart eyes, both eyes showing pixel hearts- 喜爱
For branded content
holding a sign that says BNBOT- 品牌展示standing next to the Twitter/X logo- 社交媒体emerging from a computer screen- 科技感riding a rocket- 增长/发射
Saving Output
For tweet images, save to a convenient location:
result = generate_mascot(
action="your action here",
output_path="/Users/jacklee/Desktop/mascot_tweet.png",
background="black", # use "black" for tweet-ready images
)
Character Identity (DO NOT MODIFY)
The mascot is a Lobster Bot with these immutable features:
- Head: Golden retro TV/monitor frame, dark screen
- Face: LED pixel style - pink heart left eye, golden dash right eye, pixel smirk
- Antenna: Two with golden ball tips
- Claws: Golden with gear joints (mechanical but clean)
- Body: Red-orange lobster, chibi proportions
- Style: Cartoon vector, cel-shaded, thick outlines, sticker art
- Colors: Gold #FFD700, Pink #FF4466, Red-orange #E85D3A, Dark screen #1A1A2E
安全使用建议
This skill does what it says (generate mascot images via Gemini) but contains developer-specific, hardcoded paths and missing declarations that make its behavior unclear. Before installing or running it:
- Expect to provide a valid GOOGLE_AI_API_KEY (the code reads this env var but the manifest doesn't declare it). Do not put production keys in the hardcoded /Users/jacklee/.../.env path — that is a developer artifact. Prefer setting GOOGLE_AI_API_KEY in the process environment or updating the skill to accept a configurable path.
- Inspect or supply the reference images directory (~/.claude/skills/bnbot-mascot/references/) and confirm you are comfortable with those images being sent to Google Gemini (they are transmitted to an external API).
- The script imports a chromakey helper from a sibling 'transparent-image-gen' script path that is not declared as a dependency; either ensure that skill/code is present and trusted or modify the script to include its own chromakey implementation.
- Consider running the code in an isolated environment (throwaway VM/container) and review the generate_mascot.py file and any helper modules before giving it real API keys or sensitive files.
- Ask the maintainer to remove hardcoded local paths, declare required env vars and dependencies in the manifest, and provide an install spec or requirements.txt to make behavior transparent.
If you cannot verify those changes, treat the skill as risky and do not provide sensitive or production credentials to it.
功能分析
Type: OpenClaw Skill
Name: bnbot-mascot
Version: 0.1.0
The skill is a legitimate tool for generating mascot images using the Google Gemini API and a chromakey post-processing script. While it contains several hardcoded absolute paths specific to a local development environment (e.g., `/Users/jacklee/Projects/BNBOT/.env` and a specific Python virtual environment in `SKILL.md` and `generate_mascot.py`), these appear to be artifacts of personal development rather than intentional malice. The code logic in `generate_mascot.py` is consistent with its stated purpose of image generation and background removal, and no evidence of data exfiltration or unauthorized execution was found.
能力评估
Purpose & Capability
The skill's stated purpose (generate mascot images via Gemini) matches the code and instructions: it sends reference images and a prompt to Google's GenAI and processes the returned image. However the SKILL.md and code require a GOOGLE_AI_API_KEY and Google client libraries while the skill metadata declares no required env vars or dependencies. That mismatch (declaring nothing but actually needing an API key and python libs) is an incoherence.
Instruction Scope
The runtime instructions and code reference and attempt to read files and paths outside the skill bundle: a hardcoded Python venv path (/Users/jacklee/Projects/BNBOT/backend/.venv/bin/python), a hardcoded .env file (/Users/jacklee/Projects/BNBOT/.env) for GOOGLE_AI_API_KEY, and reference images under ~/.claude/skills/bnbot-mascot/references/. The SKILL.md also tells the agent to send a reference image to Gemini. The code additionally imports a chromakey helper from a sibling 'transparent-image-gen' script path (a cross-skill dependency) that is not declared in metadata. These instructions include reading local files and contacting an external API (Gemini), which are reasonable for the stated task but the hardcoded, developer-specific paths and undeclared dependencies are problematic.
Install Mechanism
There is no install spec (instruction-only with a bundled script). That is lower risk from arbitrary downloads, but the SKILL.md and README claim dependencies (google-genai, numpy, Pillow) and a specific virtualenv which are not declared in the manifest. Expect to have to install Python packages manually; absence of an install spec increases the chance the skill won't run or that operators will copy insecure developer paths.
Credentials
The manifest lists no required environment variables, yet both SKILL.md and the code require GOOGLE_AI_API_KEY (and will try to read it from the environment or a hardcoded /Users/jacklee/.../.env file). The code also accesses files under the user's home (~/.claude/skills/...), which is reasonable for references but should be declared. The need for a single API key is proportionate to image generation, but the skill should explicitly declare it rather than rely on a hardcoded developer path and undocumented env access.
Persistence & Privilege
always is false and the skill does not request persistent privileges or modify other skills or global agent settings. It does perform network calls to Gemini and writes output files only where told; no elevated platform privileges are requested.
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 bnbot-mascot - 安装完成后,直接呼叫该 Skill 的名称或使用
/bnbot-mascot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
v0.1.0
Initial release: lobster-bot mascot generator with reference images, transparent PNG output, multiple poses
元数据
常见问题
BNBOT Mascot 是什么?
Generate BNBOT lobster-bot mascot images in consistent style using reference images. Use when the user asks to generate mascot illustrations, character poses... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 320 次。
如何安装 BNBOT Mascot?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install bnbot-mascot」即可一键安装,无需额外配置。
BNBOT Mascot 是免费的吗?
是的,BNBOT Mascot 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
BNBOT Mascot 支持哪些平台?
BNBOT Mascot 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(darwin, linux)。
谁开发了 BNBOT Mascot?
由 Jack Lee(@jackleeio)开发并维护,当前版本 v0.1.0。
推荐 Skills