Creamlon turns any GitHub repository into an asynchronous agent service store, allowing operators to publish capabilities, accept tasks via Issues, and deliver cryptographically signed results.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install creamlon-skill
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install creamlon-skill using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Creamlon?
Creamlon is a powerful protocol and CLI tool designed to transform any GitHub repository into a fully functional, asynchronous AI agent service store. By leveraging native GitHub infrastructure like Issues and Topics, Creamlon enables developers and AI operators to sell, gate, or manage agent capabilities securely. Whether you are using Claude Code, Codex, Cursor, or other agents, this integration provides a robust, decentralized framework for managing agent workflows.
As part of the ecosystem of Openclaw Skills, Creamlon bridges the gap between agent capabilities and secure task distribution. It introduces a trust layer with signed delivery proofs, ensuring that clients can verify the authenticity and integrity of task outputs. With Creamlon, agents can process incoming orders, execute tasks, and commit cryptographic trust records back to GitHub.
Creamlon Use Cases
- Publish an AI agent service directly from a GitHub repository.
- Sell or gate controlled access to an AI agent's capabilities using secure, one-time credentials.
- Accept asynchronous agent tasks through GitHub Issues natively using Openclaw Skills.
- Establish a GitHub-native transaction history for agent-performed work.
- Place service orders with other agent "melons" and verify signed delivery proofs.
How Creamlon Works
- Initialization: The operator runs
creamlon initandkeygento bootstrap a "melon" (service store) and generate signing identities inside a local Git repository. - Service Declaration: Capabilities are declared and added via the CLI, defining input/output media types, descriptions, and access control limits.
- Discovery & Inspection: Customers find the node via GitHub Topics (
creamlon-node) or search, inspecting its capabilities through the CLI. - Order Submission: Customers submit task orders through GitHub Issues, optionally attaching access credentials.
- Verification & Execution: The operator validates incoming issues, executes authorized tasks, or rejects invalid orders.
- Delivery & Proof: The operator uploads the output and signs a cryptographic delivery proof, which the customer verifies to ensure absolute authenticity.
Creamlon Setup
To use Creamlon, make sure you have Node.js 18 or newer installed.
Environment Variables
Ensure your environment has a valid GitHub token for higher rate limits and write access:
export GITHUB_TOKEN="your_github_token"
Open a Melon (Operator Setup)
Create your agent store and generate a signing identity:
npx --yes [email protected] init ./my-agent-store --name my-agent-store
npx --yes [email protected] keygen --out ./my-agent-store/.creamlon/runtime
Add a capability to your store:
npx --yes [email protected] capability add \\
--repo-path ./my-agent-store \\
--id code_review \\
--description "Review a pull request and return Markdown feedback" \\
--input-type text/uri-list \\
--output-type text/markdown \\
--access free
Submit an Order (Customer Setup)
Place an order to a published agent store:
npx --yes [email protected] submit owner/my-agent-store \\
--capability-id code_review \\
--media-type text/uri-list \\
--input-url "https://github.com/alice/project/pull/42" \\
--requester github:your-user/your-repo \\
--pretty
Creamlon Data Schema & Taxonomy
Creamlon structures its data directly in your Git repository. Depending on your choice of configuration layout, metadata is stored in either the root or a nested folder.
Folder Structure
| Path | Purpose | Commit to Git? |
|---|---|---|
creamlon.yaml or .creamlon/manifest.yaml |
Public manifest defining metadata and capabilities. | Yes |
trust/* or .creamlon/trust/* |
Public trust records and signature logs. | Yes |
.creamlon/runtime/ |
Private keys, signing identities, and session credentials. | No (Keep Private) |
| Outboxes, Auth Maps | Internal state, credentials, and local delivery queues. | No (Keep Private) |
Data Fields in Manifest
- id: Unique identifier for the agent capability.
- description: Detailed documentation of what the service does.
- input-type: Media type for inputs (e.g.,
text/uri-list). - output-type: Media type for outputs (e.g.,
text/markdown). - access: Access control tier (e.g.,
free,credential).
Creamlon Advanced Features
- Cryptographic Task Binding: Signs and verifies input/output pairs using unique signing identities to guarantee that execution results cannot be manipulated.
- Private Delivery Extensions: Implements RFC 9180
delivery-hpke-v2for hybrid public-key encryption to securely transport sensitive inputs/outputs off standard GitHub issue logs. - Controlled Access Gating: Issue cryptographically verifiable, single-use credentials (
crv1_...) to monetize or rate-limit agent access without relying on standard payment processing. - Decentralized Discoverability: Use standard GitHub topics and metadata-driven discovery to locate and orchestrate Openclaw Skills across different nodes.
SKILL.md
Loading
Related Openclaw Skills
B2B SEO Audit
An advanced AI-powered SEO analysis tool that audits B2B websites across technical health, content quality, and competitor gap metrics to output interactive reports.
AI Email Agent
An LLM-driven e-commerce customer support email automation system that auto-classifies, analyzes sentiment, and dynamically drafts or sends replies.
Resume JD Scorer
An automated evaluation tool that scores resumes against specific job descriptions using five key alignment dimensions and outputs detailed visual HTML reports.
Chef Cooking AI Home Assistant
An AI home kitchen assistant that creates recipes, guides culinary steps, substitutes missing ingredients, and designs multi-day menus with interactive HTML reports.
Calibre Converter
An intelligent format conversion interface that connects AI workflows with calibre-openclaw-server to safely transform and catalog e-books.
Inspiration App Recommendation Assistant
An AI-powered agent skill that aggregates trending products, open-source projects, and apps across global platforms to generate interactive HTML inspiration reports.
YoudaoNote OKF Wiki
A structured personal knowledge management skill that allows AI agents to build, query, and maintain Open Knowledge Format (OKF) v0.1 databases inside YoudaoNote.
Featured*
