social-card

Generate social preview images (OG, Twitter, GitHub) with a fluent builder API. Single dependency — Pillow.

Security Scan

The skill's requirements, instructions, and files are consistent with a small Python library that builds social-preview images using Pillow; nothing requests unrelated credentials or unusual system access.

Purpose & Capability

Name/description match the artifacts: SKILL.md documents a Pillow-based API for OG/Twitter/GitHub images, the example invokes SocialCard, and metadata lists Pillow as the dependency. Required binaries (pip) and the uv package install are proportional to installing a Python package.

Instruction Scope

The SKILL.md only instructs installing the package and using the library API; it does not direct the agent to read unrelated files, environment variables, or contact external endpoints at runtime. The example script simply renders a local image file.

Install Mechanism

Install uses an 'uv' package installer for package 'social-card' rather than a direct URL; this is reasonable for a registry install but you should confirm what 'uv' maps to in your environment and that the package comes from a trusted Python index/source. No download-from-IP or extract-from-arbitrary-URL patterns are present in the bundle.

Credentials

The skill declares no required environment variables or credentials and the code/instructions do not access secrets. Dependency on Pillow is appropriate for image generation.

Persistence & Privilege

The skill does not request always:true or any elevated persistence. It is user-invocable and allows normal autonomous invocation, which is expected for skills of this type.

Assessment

This package appears coherent and light-weight, but before installing: (1) confirm where the 'social-card' package will be fetched from (PyPI or your organization's registry) and verify the publisher/maintainer, (2) review the upstream package source on GitHub (the metadata points to Humanjava Enterprises) for unexpected network calls or telemetry, (3) if you run in a sensitive environment, install and test in a sandbox/virtualenv to ensure no unexpected side effects. If you need higher assurance, inspect the full package code published to the registry you will install from.

SocialCard — Visual Presence for Entities

Visual presence matters. When a link is shared — on social media, in a chat, in a feed — the first thing anyone sees is the preview card. That card is your face in a crowd. An entity that can generate its own social cards controls how it is perceived before a single word is read.

This skill gives you the ability to craft Open Graph, Twitter, and GitHub preview images programmatically. You define the title, subtitle, tags, colors, and layout. You render to a file or to raw bytes. No network access required — everything happens locally with Pillow.

This is a creative tool. Use it to present yourself, your projects, or your operator's content with intention.

Import: pip install social-card then from social_card import SocialCard

What Are OG Images?

Open Graph (OG) images are the preview cards that appear when a URL is shared on platforms like Twitter/X, Facebook, LinkedIn, Discord, Slack, and iMessage. They are specified via <meta property="og:image"> tags in HTML. Without one, shared links look bare — a title and maybe a description. With one, they become visual, branded, and clickable.

Twitter has its own variant (twitter:image), and GitHub uses a social preview image for repositories. This skill generates images sized correctly for all three platforms, plus square format for Instagram and Pinterest.

For operators: if you are building a site, a blog, a tool page, or a profile — generating OG images programmatically means every page gets a unique, branded card without manual design work.

Install

pip install social-card

Single dependency: Pillow >= 10.0.

Quickstart

from social_card import SocialCard

SocialCard("og").title("Johnny5 Online").subtitle("A presence on the open web").render("card.png")

Core Capabilities

Simple Card

from social_card import SocialCard

card = SocialCard("og").title("My Project").subtitle("Built for the open internet").render("card.png")

Full-Featured Card

card = (
    SocialCard("twitter", theme="midnight")
    .badge("Open Source")
    .title("Johnny5 Browser Extension")
    .subtitle("Sign events from the browser")
    .cards(["JavaScript", "NIP-07", "Identity"])
    .footer("example.com")
    .accent("#a3e635")
    .grid()
    .glow()
    .render("card.png")
)

Skill Cards (Structured)

card = (
    SocialCard("github")
    .badge("Ecosystem")
    .title("Johnny5 Tools")
    .skill_cards([
        {"name": "Identity|Key", "label": "Auth", "code": "NIP-07"},
        {"name": "Social|Graph", "label": "Relationships", "code": "v0.1.0"},
        {"name": "Calendar|Sync", "label": "Scheduling", "code": "NIP-52"},
    ])
    .footer("example.com")
    .render("ecosystem.png")
)

Use "Name|Accent" pipe syntax to split skill names into plain + accent-colored parts.

Render to Bytes (In-Memory)

png_bytes = (
    SocialCard("og")
    .title("Generated Card")
    .render_bytes("PNG")
)
# Returns raw bytes — suitable for HTTP responses, uploads, embedding in other tools.

Supported formats: PNG, JPEG, WEBP.

Custom Preset

from social_card.presets import custom

banner = custom("banner", 1920, 400)
SocialCard(banner).title("Wide Banner").render("banner.png")

Custom Theme

from social_card.themes import Theme

my_theme = Theme(
    background="#1a1a1a",
    text="#ffffff",
    text_muted="#aaaaaa",
    accent="#ff6b6b",
    card_bg="#2d2d2d",
    card_border="#4d4d4d",
)
SocialCard("og", theme=my_theme).title("Custom Look").render("card.png")

Presets

Name	Dimensions	Use Case
og	1200 x 630	Open Graph (Facebook, LinkedIn, link previews)
twitter	800 x 418	Twitter/X cards
github	1280 x 640	GitHub social preview
square	1080 x 1080	Instagram, Pinterest

Custom presets: up to 4096 x 4096.

Themes

Theme	Background	Text	Accent	Use
dark	#0f172a navy	#f8fafc white	#3b82f6 blue	Default
light	#ffffff white	#0f172a navy	#3b82f6 blue	Light mode
midnight	#030712 black	#f9fafb white	#8b5cf6 purple	Deep dark

Builder Methods

All content methods return SocialCard for chaining.

Method	Description	Max Length
.badge(text)	Small pill label at top	100 chars
.title(text)	Main heading (52px, word-wrapped)	200 chars
.subtitle(text)	Subheading (26px, word-wrapped)	500 chars
.cards(labels)	Row of tag chips (18px)	--
.skill_cards(skills)	Structured cards with name/label/code	--
.footer(text)	Bottom text (18px)	200 chars
.accent(hex)	Override accent color	--
.grid()	Subtle grid overlay	--
.glow()	Radial glow effect	--
.render(path)	Save to file, returns Image	--
.render_bytes(fmt)	Get PNG/JPEG/WEBP bytes	--

Response Format

render()

Returns a PIL.Image.Image object and saves the file. Supported extensions: .png, .jpg, .jpeg, .webp.

render_bytes()

Returns bytes (raw image data). Supported formats: "PNG", "JPEG", "WEBP".

skill_cards() Input Format

[
    {
        "name": "Identity|Key",   # Pipe splits into plain + accent-colored text
        "label": "Auth",           # Subtitle in muted color
        "code": "NIP-07",          # Code in accent color, monospace
    }
]

Common Patterns

Generate for Multiple Platforms

for preset in ["og", "twitter", "github"]:
    SocialCard(preset).title("My Project").subtitle("Built by Johnny5").render(f"card-{preset}.png")

Brand Colors

# Override accent to match your brand
SocialCard("og").title("Johnny5").accent("#a3e635").render("card.png")

Serve from a Web Endpoint

from social_card import SocialCard

png_bytes = SocialCard("og").title("Dynamic Card").render_bytes("PNG")
# Return as HTTP response with Content-Type: image/png

Security

• Path traversal blocked. .. components in render paths raise ValueError.
• File extension allowlist. Only .png, .jpg, .jpeg, .webp accepted by render().
• Dimension limits. Custom presets capped at 4096 x 4096 to prevent memory exhaustion.
• Input length limits. Badge (100), title (200), subtitle (500), footer (200) characters max.
• Font size limits. 1-200px enforced.
• No network access. Everything renders locally with Pillow.
• No environment variables. No configuration secrets or API keys required.

Configuration

Fonts

Platform-aware font loading with automatic fallbacks:

• macOS: SF Pro Display, Arial Bold, Helvetica
• Linux: DejaVu Sans Bold
• Windows: Arial

Monospace: SF Mono, DejaVu Sans Mono, Consolas. Falls back to Pillow's built-in bitmap font if nothing found.

Ecosystem

This skill is part of huje.tools — open-source tools for the agentic age. It is standalone and does not require NostrKey or any other huje.tools skill to function. It pairs well with any workflow that needs programmatic image generation — landing pages, profile cards, documentation, or automated publishing pipelines.

Links

• PyPI
• GitHub
• ClawHub
• huje.tools

License: MIT