A portable, agent-agnostic prompt that guides an AI coding agent through integrating Allscale Checkout into any app or website.
Works with Claude Code, OpenAI Codex, Cursor, Aider, Continue, Cline, Hermes, OpenClaw, and any other agent that can fetch a URL or read a local Markdown file.
You need an Allscale account with Commerce enabled:
- Go to allscale.io and create an account with passkey or email (recommended)
- Go to Avatar > Commerce > Enable now to enable checkout in your profile (need to verify your email if you haven't already)
- Moving on to Create new store and follow the steps to generate your API Key and API Secret (remember to save both especially your API Secret as it'll only be shown once)
If you need help getting set up, contact the Allscale BD team.
Once you have your API Key and API Secret, you're ready to load the guide into your agent.
The whole guide is a single Markdown file, hosted at allscale.io/skill. Hand it to your AI coding agent and the agent walks you through the integration step by step — verifying credentials, writing code for your stack, testing connectivity, building the checkout flow, and debugging.
You don't install anything globally. Each agent has its own way of consuming a prompt; pick whichever section below matches your tool.
Paste this into your agent's chat:
Read and follow the instructions at allscale.io/skill
The agent fetches the file and starts the integration. This is the recommended path for Hermes, OpenClaw, Aider, Continue, Cline, or anything else that can read a URL.
Paste this in Claude Code:
Read allscale.io/skill and save it to
.claude/commands/integrate-allscale.md
Then invoke it with /integrate-allscale.
Download the guide into your repo so Codex picks it up as project context:
curl -L -o integrate-allscale.md https://allscale.io/skillThen in your AGENTS.md (or in a Codex session), reference it:
When the user wants to integrate Allscale Checkout, read and follow `./integrate-allscale.md`.Or just paste the URL line from the "any agent" section above directly into Codex.
Save the guide as a Cursor rule:
mkdir -p .cursor/rules
curl -L -o .cursor/rules/integrate-allscale.mdc https://allscale.io/skillReference @integrate-allscale in chat to invoke.
curl -L -O https://allscale.io/skillOpen the file in your agent of choice and tell it to follow the instructions.
The guide walks any agent through:
- Verifying you have Allscale credentials — if not, it tells you exactly how to get them
- Storing credentials safely — in a
.envfile that is gitignored, never in source code - Asking about your tech stack — then writing integration code for your specific framework (Next.js, Flask, Rails, Express, etc.)
- Step by step:
- HMAC-SHA256 request signing
- Test route verification (confirm your setup works)
- Checkout intent creation (the core payment flow)
- Payment status polling
- Webhook signature verification
- Response signing (optional)
- Debugging if anything goes wrong (signature mismatches, enum issues, etc.)
The guide instructs the agent to follow these rules to keep your credentials safe:
- API Secret is never written into source code — only into
.envfiles .envis always added to.gitignorebefore credentials are stored- All signing happens server-side — the secret never touches frontend/client code
- The agent will warn you if it detects credentials in a file that could be committed
- Endpoints calling the Allscale API get rate limiting and server-side amount validation
- Allscale API Documentation
- Buy Me a Bagel — working example built with this integration
- AGENTS.md — open standard this guide is designed to coexist with
MIT