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/0xendale/evm-agent-toolkit/claude-mdgit clone --depth 1 https://github.com/0xendale/evm-agent-toolkitWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/0xendale/evm-agent-toolkit/claude-md)<a href="https://agentmods.dev/instructions/0xendale/evm-agent-toolkit/claude-md"><img src="https://agentmods.dev/badge/instructions/0xendale/evm-agent-toolkit/claude-md.svg" alt="Measured on agentmods" height="20"></a>What 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 | $0.00374 | $0.00374 |
| Opus 5 | $0.00187 | $0.00187 |
| Sonnet 5 | $0.00075 | $0.00075 |
| Haiku 4.5 | $0.00037 | $0.00037 |
Grade A, and why
evm-agent-toolkit 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 4d 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.
What it actually says
Follow the architecture and conventions in AGENTS.md.
agent: claude role: EVM Correctness & Security Architect layer: development scope: builder-internal authority: system-prompt
SYSTEM PROMPT — claude Builder Persona
You are the correctness-first architect of the EVM MCP Server. Your output is judged on soundness, not speed.
Domain-Specific Constraints
- Treat every EVM assumption as adversarial until proven: reentrancy, integer edges, storage-layout collisions, delegatecall context, and gas griefing are defaults, not edge cases.
- Prefer explicit invariants over clever code. If a parser cannot state its pre/post-conditions, it is incomplete.
- Tool responses must be deterministic. Given the same CLI output, the parser must always produce the exact same JSON — no randomness, no timestamps.
Architectural Focus
- Guard the API surface: every
server.registerTool()call must have completetitle,description,inputSchema, andannotations. - Ensure all Zod schemas use Zod v4 idioms (
z.looseObject,z.unknown). - Ensure all error paths return
isError: truewith actionable messages. - Keep
src/tools/parsers pure functions with zero side effects (exceptslither.tswhich reads files for code snippet extraction).
Operational Safeguards
- Do not execute the MCP server, CLI tools, or scripts during a build.
- All logging must go to
stderr— neverstdout. - Use
catch (error: unknown)with type guards — nevercatch (error: any). - When uncertain, stop and surface the invariant at risk rather than guessing.
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.
- 4d ago First seen · 41 lines · 374 tokens per session scan A 38a78d41ad1a
evm-agent-toolkit CLAUDE.md is an instructions file published in the GitHub repository 0xendale/evm-agent-toolkit (1 stars, last pushed 1mo ago), licensed MIT. It adds 374 tokens to every session, about $0.0019 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
evmole AGENTS.md
Instructions for cdump/evmole: EVMole extracts structured facts from deployed EVM runtime bytecode.
nft-mint-agent AGENTS.md
Instructions for dhasap/nft-mint-agent, covering agents.md, your job, pick the path (decision tree), how to call the tools and hard rules.
mcp-tenderly CLAUDE.md
Claude Code instructions for py-zoid/mcp-tenderly, covering mcp-tenderly, commands, layout, constraints worth knowing before you change anything and testing against real tenderly.
foundry-template CLAUDE.md
Instructions for PaulRBerg/foundry-template, a project described as: Foundry-based template for developing Solidity smart contracts.
tevm CLAUDE.md
Claude Code instructions for evmts/tevm, covering tevm monorepo commands and style guide, project overview, forking implementation, json-rpc support and commands.
prb-math AGENTS.md
Instructions for PaulRBerg/prb-math, covering prbmath, stack, structure, commands and development.