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/dromara/liteflow/claude-mdgit clone --depth 1 https://github.com/dromara/liteflowWhat 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 | $0.04675 | $0.04675 |
| Opus 5 | $0.02337 | $0.02337 |
| Sonnet 5 | $0.00935 | $0.00935 |
| Haiku 4.5 | $0.00468 | $0.00468 |
Grade A, and why
liteflow 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 3d 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 — 266 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
本文件为 Claude Code (claude.ai/code) 在本仓库中工作时提供指引。
概述
LiteFlow(v2.16.1)是一个轻量级规则引擎框架,用于复杂的组件化业务编排。它通过 DSL 驱动工作流,支持热重载和 11 种脚本语言。项目面向 Java 8+(最高支持到 JDK 25),拥有 2000+ 测试用例。
官方文档:https://liteflow.cc/pages/5816c5/
命令
构建
# 构建整个项目(模块集由 JDK 版本自动选择 —— 见下方说明)
mvn clean package -DskipTests
# 带测试构建
mvn clean package
# 构建指定模块
mvn clean package -DskipTests -pl liteflow-core
# 发布生产模块(手动选择 profile,不会自动激活)
mvn clean package -DskipTests -P release-on-8 # JDK 8 发布产物
mvn clean package -DskipTests -P release-on-17 # react-agent + spring-boot4-starter
JDK 驱动的模块选择(重要): 根 pom.xml 中没有顶层的 <modules> 块 —— 构建的模块列表完全由按 JDK 激活的 profile 提供:
compile-8-to-16(在 JDK[1.8,17)上激活):core、scripts、rules、spring-boot-starter、spring、solon、testcase-el、el-builder、metrics、benchmark。compile-17+(在 JDK[17,)上激活):以上全部,外加liteflow-react-agent和liteflow-spring-boot4-starter。
因此 liteflow-react-agent 和 liteflow-spring-boot4-starter 只有在 JDK 17+ 上构建时才会被纳入。这两个模块以 maven.compiler.target=17 编译,但根 pom 注释指出 agentscope-java(react-agent 背后的引擎)运行时需要 Java 21+ —— 本仓库当前的开发 JDK 为 21。release-on-8 / release-on-17 这两个 profile 没有 <activation>,必须用 -P 手动选择:release-on-8 发布 JDK 8 产物(含 liteflow-metrics),release-on-17 只发布 liteflow-react-agent 与 liteflow-spring-boot4-starter。
运行测试
测试用例的编写位置(强制约定): 所有测试用例必须写在
liteflow-testcase-el/下对应的测试模块中,禁止写在liteflow-core、liteflow-metrics、liteflow-spring*等核心代码模块下。核心代码模块只放生产代码,不放测试。新增功能时,请在liteflow-testcase-el/下选择(或新建)合适的测试模块来覆盖该功能(按框架 / 配置源 / 脚本 / 特性维度划分,见下文「测试基础设施」)。
注意: 根 pom 的 surefire 默认
skipTests=true,直接mvn test不会真正执行测试,需要显式加-DskipTests=false。
# 运行所有测试
mvn test -DskipTests=false
# 运行指定模块的测试(liteflow-testcase-el/ 下有 30+ 个测试模块)
mvn test -DskipTests=false -pl liteflow-testcase-el/liteflow-testcase-el-springboot
# 运行单个测试类
mvn test -DskipTests=false -pl liteflow-testcase-el/liteflow-testcase-el-springboot -Dtest=FlowExecutorTest
# 运行指定的测试方法
mvn test -DskipTests=false -pl liteflow-testcase-el/liteflow-testcase-el-springboot -Dtest=FlowExecutorTest#testExecute
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.
- 3d ago First seen · 266 lines · 4,675 tokens per session scan A 32703b4f6f9a
liteflow CLAUDE.md is an instructions file published in the GitHub repository dromara/liteflow (3,842 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 4,675 tokens to every session, about $0.0234 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.
Other instructions, from other repositories
Blaze CLAUDE.md
Claude Code instructions for grloper/Blaze, covering claude.md, what this is, commands, architecture (the big picture) and invariants you must not break.
java_upgrade CLAUDE.md
Instructions for kousen/java_upgrade, covering project context for claude code, project structure, key features, recently added exercises and java features covered.
sivalabs-agent-skills AGENTS.md
Instructions for sivaprasadreddy/sivalabs-agent-skills, covering agent instructions, installation methods, method 1: claude code plugin installation, add the agent-skills marketplace and install plugins.
Ruvyxa AGENTS.md
Instructions for thirawat27/Ruvyxa, covering ruvyxa monorepo agent guide, repository shape, operating rules, the child-process stdio protocol and cache identity is derived, never stamped.
markdy-com copilot-instructions.md
Instructions for HoangYell/markdy-com, a project described as: 🎬 Open-source animated architecture diagrams as code. Diagram-native DSL, 17 layout engines, Web Animations API, and MCP server for AI agents.
quarkus-agentic-scaffolding AGENTS.md
Instructions for eldermoraes/quarkus-agentic-scaffolding, covering quarkus + langchain4j + ai stack - project conventions, version: 0.21.0, 1. required tooling (mandatory), 2. java conventions and 3. quarkus conventions.