What is Sherblock CLAUDE.md?

A project instruction file for Sherblock, a system that analyzes blockchain transactions and addresses using several cooperating AI agents. It explains the project structure, Windows setup, and how the agents coordinate work.

Use it when modifying Sherblock’s question, planning, execution, or orchestration agents and their JavaScript environment.

It gives a coding agent the project’s architecture and operating rules so changes fit the existing system.

Sherblock CLAUDE.md — Instructions file by scriptLin-bjtu

Install

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.

agentmods

npx agentmods add instructions/scriptlin-bjtu/sherblock/claude-md

Clone the repo

git clone --depth 1 https://github.com/scriptLin-bjtu/Sherblock

Per session 4,472 This file is loaded in full into every session.

When invoked 4,472 The same file — it is already loaded in full.

Security scan A 0 findings. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

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	$0.04472	$0.04472
Opus 5	$0.02236	$0.02236
Sonnet 5	$0.00894	$0.00894
Haiku 4.5	$0.00447	$0.00447

Measured 2d ago against content hash 20a80d914ab9, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

Sherblock 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.

CLAUDE.md · 406 lines

How it starts

The opening of the file, as written. The whole thing — 406 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Project Overview

This is a Blockchain Transaction Behavior Analysis Agent - a multi-agent system for analyzing blockchain transactions and addresses. It uses a Plan-and-Execute architecture with three specialized agents coordinated by a central orchestrator. The project uses ES Modules ("type": "module").

QuestionAgent (src/agents/questionBot/) - Interactive information collection using ReAct pattern
PlanAgent (src/agents/planBot/) - Strategic planning using deepseek-v4-flash thinking mode
ExecuteAgent (src/agents/executeBot/) - Step execution using ReAct pattern with blockchain skills
AgentOrchestrator (src/agents/orchestrator/) - Central coordinator managing workflow state transitions

System Environment

This system runs on Windows platform. All file paths, shell commands, and environment configurations should be compatible with Windows operating system.

Architecture

Agent Architecture (Plan-and-Execute)

User Input → QuestionAgent → PlanAgent → ExecuteAgent (per step)
                                              ↓
                                         Blockchain Data

Workflow Stages (managed by WorkflowStateMachine):

IDLE → COLLECTING → PLANNING → EXECUTING → REVIEWING → COMPLETED

Parallel Execution Mode

The system supports both serial and parallel execution modes:

Serial mode (default): Steps execute sequentially in a while loop
Parallel mode: Steps execute concurrently using DAG-based dependency resolution

Supported since commit: cd83e4b (Merge branch 'feature/parallel-tasks')

Key differences:

Feature	Serial Mode	Parallel Mode
Plan structure	steps array	DAG (Directed Acyclic Graph)
Execution	while loop	p-limit concurrency
Dependencies	next_step_hint	explicit depends_on + implicit inference
Scope updates	direct write	lock mechanism

Read the full file on GitHub · 406 lines

Changes

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 · 406 lines · 4,472 tokens per session scan A 20a80d914ab9

Subscribe to this mod's changes

Sherblock CLAUDE.md is an instructions file published in the GitHub repository scriptLin-bjtu/Sherblock (2 stars, last pushed 2mo ago), licensed MIT. It adds 4,472 tokens to every session, about $0.0224 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.

agent ai-agents blcokchain

Sherblock CLAUDE.md

Getting it into your agent

What it costs to keep this loaded

Grade A, and why

Nothing flagged

How it starts

CLAUDE.md

Project Overview

System Environment

Architecture

Agent Architecture (Plan-and-Execute)

Parallel Execution Mode

What this file has done since we first saw it

Other instructions, from other repositories

ii-agent AGENTS.md

intelligent-terminal rust.instructions.md

InvestSkill GEMINI.md

sortie copilot-instructions.md

crust AGENTS.md

she-love-me CLAUDE.md