Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/johnaverse/chains-api/claude-mdgit clone --depth 1 https://github.com/Johnaverse/chains-apiWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.04601 | $0.04601 |
| Opus 5 | $0.02301 | $0.02301 |
| Sonnet 5 | $0.00920 | $0.00920 |
| Haiku 4.5 | $0.00460 | $0.00460 |
Grade A, and why
chains-api CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 259 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md — Chains API
Project Overview
Chains API is a Node.js service that aggregates blockchain chain data from five external sources, maintains an in-memory index, and exposes it via a REST API (Fastify), MCP stdio server, and MCP HTTP server. No database — data is fetched from remote JSON/Markdown sources, indexed in memory, and optionally cached to disk for stale-first startup.
Single data source per container: server.js runs the REST API and the MCP HTTP server in one process (default npm start / Docker CMD). Both surfaces read the same module-level in-memory store (src/store/), and only the REST side starts the refreshers — so a container never refreshes the same public RPC endpoint twice. index.js (REST only) and mcp-server-http.js (MCP only) remain runnable standalone for local/dev use. The MCP tool surface in mcp-tools.js mirrors the REST read endpoints (chains, search, relations, endpoints, clients, scaling, slip44, status-pages, stats, keywords, rpc-monitor, refresher, validate, live-incidents, forum-news).
Upstream feed services (separate deployables): live incidents and forum news do not come from the five data sources — they are served by two sibling Node services with their own repos, releases, and containers:
| Service | Feeds | Consumed via |
|---|---|---|
chains-status-news |
Chain + RPC-provider status pages → incidents, scheduled maintenance | LIVE_INCIDENTS_URL |
chains-forum-news |
Official community/governance forums, keyed by chain ID | FORUM_NEWS_URL |
The wire contract shared by all three services — addressing, health/readiness, versioning, collection envelopes, error shape, WebSocket framing, timestamps, auth — is specified in docs/SERVICE-CONTRACT.md. Read it before changing any cross-service payload.
Both are independent git repos, so they are often checked out next to this one in the same working tree (chains-status-news/, chains-forum-news/) — useful for cross-checking a feed's payload against a consumer here. They are versioned and released separately from this project, and the paths in this document, in package.json, and in .github/workflows/ are all relative to this project's root.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 259 lines · 4,601 tokens per session scan A 3f506209fb4d
chains-api CLAUDE.md is an instructions file published in the GitHub repository Johnaverse/chains-api (5 stars, last pushed 4d ago), licensed MIT. It adds 4,601 tokens to every session, about $0.0230 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
explorer AGENTS.md
AGENTS.md instructions for aptos-labs/explorer, covering aptos explorer - agent guidelines, quick reference, project structure, routing and navigation and how routes are defined.
explorer copilot-instructions.md
Copilot instructions for aptos-labs/explorer, a project described as: The explorer of the Aptos blockchain.
explorer CLAUDE.md
Claude Code instructions for aptos-labs/explorer, a project described as: The explorer of the Aptos blockchain.
BNB Chain MCP Agent Guide
AGENTS.md instructions for nirholas/bnbchain-mcp, covering 🤖 bnb chain mcp - ai agent integration guide, 🎭 agent personas, 1. defi analyst agent, 2. portfolio manager agent and 3. security auditor agent.
bnbchain-mcp copilot-instructions.md
Copilot instructions for nirholas/bnbchain-mcp, covering bnbchain-mcp and terminal management.
bnbchain-mcp CLAUDE.md
Claude Code instructions for nirholas/bnbchain-mcp, covering bnbchain-mcp and terminal management.