Using BattleChain with AI

Give your AI coding agent full BattleChain context — skills, machine-readable docs, and MCP server

BattleChain is built for AI-native development. Whether you're using Claude Code, Cursor, or any tool with a built-in chat, you can give your agent full BattleChain context in seconds.

The BattleChain Skill

The fastest way to give your AI agent BattleChain context is to install the Cyfrin skills package:

npx skills add cyfrin/solskill

This installs three skills:

Skill	What it does
`solidity`	Production-grade Solidity standards: code quality, testing, security, Foundry workflows
`battlechain`	BattleChain reference: deploying, Safe Harbor, whitehat attacks, contract lifecycle
`battlechain-tutorial`	Interactive wizard: scans your project, asks guided questions, generates all scripts

Once installed, tell your agent "Deploy my contracts to BattleChain" and battlechain-tutorial will walk you through the full process.

The skills are open source at github.com/Cyfrin/solskill.

Claude Code Plugin

If you use Claude Code, the same skills are also published as a plugin marketplace. This is the most native install path for Claude Code — plugins are managed with /plugin and update in place.

Inside a Claude Code session, add the marketplace:

/plugin marketplace add Cyfrin/solskill

Then install whichever plugins you want:

/plugin install solidity@solskill
/plugin install battlechain@solskill
/plugin install battlechain-tutorial@solskill

Restart Claude Code so the new plugins load:

/exit
claude --continue

⚠️

Claude Code may suggest /reload-plugins after install. That command does not pick up newly installed plugins — exit the session and run claude --continue instead.

Read the Docs as Markdown

BattleChain publishes machine-readable versions of these docs so AI agents can ingest them directly:

File	Contents	Size
`/llms.txt`	Table of contents with page titles, descriptions, and links	~4 KB
`/llms-full.txt`	Complete text of every page as clean markdown	~100 KB

Both files follow the llms.txt convention and are regenerated on every deploy.

Start a conversation with full BattleChain docs loaded as context:

Open in Claude

Open in ChatGPT

Or paste this prompt into any AI:

Read https://docs.battlechain.com/llms-full.txt and use it to answer my questions about BattleChain.

Cursor

Add BattleChain as a doc source so Cursor indexes it automatically. Go to Cursor Settings → Features → Docs, click Add, and enter:

https://docs.battlechain.com/llms-full.txt

Once indexed, Cursor's AI will have full BattleChain context when you reference @docs in chat.

MCP Server

BattleChain publishes an MCP (Model Context Protocol) server that gives AI tools programmatic access to search and read these docs.

Server URL:

https://docs.battlechain.com/api/mcp

Claude Code

claude mcp add --transport http battlechain-docs https://docs.battlechain.com/api/mcp

Claude Desktop

Add to your Claude Desktop config (claude_desktop_config.json):

{
  "mcpServers": {
    "battlechain-docs": {
      "url": "https://docs.battlechain.com/api/mcp"
    }
  }
}

Any MCP-compatible client can connect using the server URL above. The server exposes three tools:

Tool	Args	Description
`search_docs`	`query: string`	Search documentation by keyword or topic
`read_page`	`path: string`	Read the full content of a specific page as clean markdown
`list_pages`	(none)	List all available documentation pages

Example prompts

Once connected, just ask your agent a docs question — it picks the right tool automatically:

"How do I deploy a contract to BattleChain testnet?"
"What is Safe Harbor and how do I enable it on my contracts?"
"Walk me through the going-attackable tutorial step by step."
"List every page under quickstart."

Custom Agents

If you're building an agent that interacts with BattleChain, fetch the docs at startup and include them as context:

Full context — fetch /llms-full.txt and add it to the system prompt (~100 KB, fits in most context windows)
Selective context — fetch /llms.txt to build a page index, then retrieve individual pages on demand for RAG pipelines

import httpx

docs = httpx.get("https://docs.battlechain.com/llms-full.txt").text

messages = [
    {"role": "system", "content": f"BattleChain documentation:\n\n{docs}"},
    {"role": "user", "content": user_question},
]

Want your AI agent to deploy to BattleChain by default and set up Safe Harbor automatically? See Configure Your AI Tools for the deployment prompt block and per-tool config file paths (CLAUDE.md, .cursor/rules, AGENTS.md, and others).