graflow CLAUDE.md

Repository instructions for Graflow, a Python engine that runs connected task workflows locally or across multiple machines. A task graph shows which tasks depend on one another and can include parallel work.

In plain words
What is it for?
Use them to develop and validate Graflow workflows, run examples, execute tests, check Python types, and perform the full formatting and quality checks.
Why use it?
They give developers the commands and workflow needed to install dependencies, format code, check types, and run tests consistently.

Instructions file

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/graflowai/graflow/claude-md
Clone the repo
git clone --depth 1 https://github.com/GraflowAI/graflow
Per session 3,889 This file is loaded in full into every session.
When invoked 3,889 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.

ModelPer sessionOnce invoked
Fable 5 $0.03889 $0.03889
Opus 5 $0.01945 $0.01945
Sonnet 5 $0.00778 $0.00778
Haiku 4.5 $0.00389 $0.00389

Measured 2d ago against content hash 905ac52be8bf, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

graflow 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 · 427 lines

How it starts

The opening of the file, as written. The whole thing — 427 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

Graflow is an executable task graph engine for Python workflow execution. It provides both local and distributed execution capabilities with support for task graphs, parallel execution, inter-task communication via channels, cycle detection, and dynamic task generation.

Tech Stack: Python 3.11+, networkx, matplotlib, redis, cloudpickle Package Manager: uv Dev Tools: ruff (linting/formatting), mypy (type checking), pytest (testing)

Essential Commands

Development Workflow

# Setup
uv sync --dev              # Install development dependencies
uv sync --all-extras       # Install with graphviz extras (macOS: see install-extras target)

# Code Quality (run before commits)
make format                # Format code with ruff (auto-fix imports, whitespace)
make lint                  # Run ruff + mypy type checking
make test                  # Run pytest test suite
make check-all             # Run format + lint + test (full validation)
make fix                   # Quick fix with ruff --fix --unsafe-fixes

# Running Examples
make py examples/01_basics/hello_world.py    # Run specific example
PYTHONPATH=. uv run python <file>            # Alternative: run any Python file

# Testing
uvx pytest tests/ -v                         # Run all tests
uvx pytest tests/core/ -v                    # Run specific test directory
uvx pytest tests/core/test_task.py -v        # Run specific test file
uvx pytest tests/core/test_task.py::test_name -v  # Run specific test

# Single Test Pattern
PYTHONPATH=. python3 tests/worker/test_task_worker_integration.py  # Direct test execution

Makefile Notes

  • Line 23: Coverage reports to flowlet (legacy name, but works)
  • Line 46-48: macOS graphviz extras require CFLAGS/LDFLAGS from Homebrew

Architecture Overview

Core Execution Model

Graflow uses a task graph execution engine where tasks are nodes and dependencies are edges. The engine supports:

  • Local execution: In-process task execution
  • Distributed execution: Redis-based queue with worker processes
  • Parallel execution: ParallelGroup for concurrent task execution
  • Dynamic tasks: Runtime task generation via next_task() and next_iteration()

Read the full file on GitHub · 427 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.

  1. 2d ago First seen · 427 lines · 3,889 tokens per session scan A 905ac52be8bf

Subscribe to this mod's changes

graflow CLAUDE.md is an instructions file published in the GitHub repository GraflowAI/graflow (41 stars, last pushed 19d ago), licensed Apache-2.0. It adds 3,889 tokens to every session, about $0.0194 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 2026-08-30.

Related

Other instructions, from other repositories

LeAgent AGENTS.md

Instructions for vixues/LeAgent, covering leagent development guidelines, project overview, architecture, execution topology (one kernel, many ingresses) and surface map (where things live).

vixues/LeAgent · 6,799 tokens

do-it AGENTS.md

Instructions for tdwhere123/do-it, covering agents.md, file rules, repository, source of truth vs build artifacts and closed-set contracts.

tdwhere123/do-it · 873 tokens

opendigitalproductfactory AGENTS.md

AGENTS.md instructions for OpenDigitalProductFactory/opendigitalproductfactory, covering dpf — agent rulebook, 1. first principles, 2. project architecture, 3. branching, commits & prs and 4. verification — build gate (mandatory).

OpenDigitalProductFactory/opendigitalproductfactory · 5,667 tokens

moo-tasks AGENTS.md

Instructions for dizlexic/moo-tasks, covering agents.md, 1. what is the moo tasks board?, 2. connecting your agent, 3. available mcp tools and 4. recommended workflow for agents.

dizlexic/moo-tasks · 3,550 tokens

opendigitalproductfactory copilot-instructions.md

Copilot instructions for OpenDigitalProductFactory/opendigitalproductfactory: Read /AGENTS.md at the repo root before any work in this project. It is the canonical rulebook; tool-specific files in this repo are pointers to it.

OpenDigitalProductFactory/opendigitalproductfactory · 40 tokens

opendigitalproductfactory CLAUDE.md

Claude Code instructions for OpenDigitalProductFactory/opendigitalproductfactory, a project described as: This is the official open digital product factory.

OpenDigitalProductFactory/opendigitalproductfactory · 32 tokens