- Home
- Skills
- Agents & Orchestration
- stylelock
More screenshots
Works with the AI Agents you already use
stylelock
Enforce visual consistency for AI image generation using versioned style guides, reference anchors, and CLI provenance.
$5
/moSubscription · cancel anytime
Secure checkout via Stripe
- Always the latest version
- Delivered live to your agent via MCP
- Cancel anytime, access ends at period end
stylelock
Example session with this skill installed
Set up a new project in the current directory and generate 2 transparent 'potion bottle' icons in our pixel art style.
- Read your context and instructions
- Compiled the stylelock
I have initialized the stylelock project and generated the requested assets using the pixel-art-v1 style guide.
Generation Results
| Asset ID | Subject | Type | Status | Path |
|---|---|---|---|---|
| img_8f23 | Potion bottle | object | pending_review | ./art/gen/potion_01.png |
| img_9a41 | Potion bottle | object | pending_review | ./art/gen/potion_02.png |
// Sidecar excerpt for potion_01.png.json
{
"style_version": "1.0.2",
"provider": "openai",
"model": "dall-e-3",
"anchors": ["ref_pixel_hero", "ref_pixel_tree"],
"prompt_metadata": {
"subject": "potion bottle",
"style_override": "hard_rules: no_anti_aliasing, max_colors_16"
}
}
Next steps
- Run
stylelock review pendingto inspect the renders. - Use
scripts/strip_contact_shadow.pyif the bottles have baked-in shadows. - Run
stylelock review accept ./art/gen/potion_01.pngif the style match is correct.
Subscription · always the latest version
- Secure checkout via Stripe
- Cancel anytime
- Delivered live via MCP
About this skill
Maintaining visual consistency across an AI generation workflow is notoriously difficult. Prompting alone often drifts, leading to assets that feel disconnected from your game or application brand. This skill provides a formal interface to stylelock, a CLI tool that enforces strict style guides and reference anchors on every image call.
What it does
- Enforce Style Consistency through versioned style guides that override model-specific biases.
- Reference Anchoring to lock in specific palettes, textures, or compositions using existing art as ground truth.
- Automated Provenance for tracking every generation with sidecar metadata including the exact prompt and style version used.
- Image Post-Processing via built-in scripts for chroma keying, shadow removal, and character sheet splitting.
- Style Governance including linting, validation, and immutable versioning (style bumping) to prevent drift.
How it works
- Initialize a project using
stylelock initto establish a local workspace and configuration. - Build the guide by registering 3-10 reference images and measuring their technical properties (palette, rendering).
- Draft the spec in
style-guide-v1.jsonto define "hard rules" like text exclusion or specific lighting constraints. - Execute generations via
stylelock generate, which injects the style spec and anchors into the model call. - Review and refine using the
accept/rejectworkflow to feed rejection reasons back into the style guide.
Frameworks & tools
The skill requires a Python environment with uv or pipx. It utilizes the Pillow library for local image manipulation and supports providers like openai, gemini, and replicate.
Why this beats prompting it yourself
Relying on ad-hoc prompts leads to "style bleed" where the model's default training data gradually dilutes your house style. This tool decouples the "what" (subject) from the "how" (style), ensuring that even as you scale to hundreds of assets, the visual language remains identical.
Use cases
- Generating a 200-item icon set that must match a specific vector aesthetic.
- Creating background plates and transparent sprites that fit a single game world.
- Maintaining character consistency across multiple poses and environments.
- Building a multi-provider pipeline that produces similar results regardless of the underlying LLM.
Known limitations
Does not support direct API calls; all generations must go through the CLI to maintain the provenance record. Requires a local Python installation for the helper scripts and CLI.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
Requires local Python environment and uv/pipx. All image generations must go through the stylelock CLI. Supports specific providers (OpenAI, Gemini, Replicate).
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed today
- Software
Python