← 返回 Skills 市场
Ralph Mode - Autonomous Development Loops
作者
richginsberg
· GitHub ↗
· v1.2.0
3428
总下载
6
收藏
10
当前安装
3
版本数
在 OpenClaw 中安装
/install ralph-mode
功能描述
Autonomous development loops with iteration, backpressure gates, and completion criteria. Use for sustained coding sessions that require multiple iterations, test validation, and structured progress tracking. Supports Next.js, Python, FastAPI, and GPU workloads with Ralph Wiggum methodology adapted for OpenClaw.
安全使用建议
This skill appears coherent and does not request secrets or install arbitrary code, but it coordinates broad repository operations: it will read project files, run build/test/typecheck/lint commands, and instruct sub-agents to implement and commit changes. Before using it: (1) run it in a sandbox or branch first so commits and commands cannot affect production; (2) review AGENTS.md and IMPLEMENTATION_PLAN.md to ensure validation commands are safe and non-destructive; (3) restrict autonomous invocation if you don't want the agent to spawn many sub-agents or make commits without manual approval; and (4) do not grant repository, CI, cloud, or GPU credentials unless you explicitly trust the workflow and have audited the commands it will run.
功能分析
Type: OpenClaw Skill
Name: ralph-mode
Version: 1.2.0
The OpenClaw AgentSkills skill bundle 'ralph-mode' is designed for autonomous, iterative software development. All files, including the `SKILL.md` instructions, `scripts/loop.sh`, and reference documentation, consistently promote structured workflows, validation gates (tests, lint, typecheck), clear progress logging, and explicit error handling. There is no evidence of malicious intent such as data exfiltration, unauthorized command execution, persistence mechanisms, or prompt injection attempts designed to bypass user intent or perform harmful actions. Instead, the instructions emphasize transparency, control, and best practices for AI agent operation, making it a benign development tool.
能力评估
Purpose & Capability
The name and description promise autonomous development loops; the SKILL.md, references, and loop.sh implement a project-first loop/coordination methodology (plans, gates, spawn sub-agents, run tests/linters/commits). There are no unrelated environment variables, binaries, or install steps that don't match the stated purpose.
Instruction Scope
The instructions direct agents and sub-agents to read project files (IMPLEMENTATION_PLAN.md, AGENTS.md, specs/, src/), run project validation commands (tests, lint, typecheck, build) and make commits. That is coherent with the stated purpose, but it grants the agent broad discretion over repository files and lifecycle actions (spawn sub-agents, run arbitrary project commands, update/commit plan files). Also the references mention spawning large numbers of sub-agents (e.g., 'up to 250 parallel Sonnet subagents'), which widens runtime activity and resource usage—this is expected for an autonomous loop tool but worth noting before granting full autonomy.
Install Mechanism
No install specification is provided (instruction-only plus a small bash helper script). Nothing is downloaded or written by an installer step beyond normal use of loop.sh and editing project files. This is the lowest-risk install posture.
Credentials
The skill declares no required environment variables, credentials, or config paths. The SKILL.md references running standard project commands (npm, pytest, mypy, etc.) but does not request secrets. This is proportionate to a coordination/methodology skill. Note: to actually push commits, run CI, or access cloud resources in real workflows a user would need to grant external credentials outside this skill—those are not requested by the skill itself.
Persistence & Privilege
always:false and no install means this skill does not demand permanent forced inclusion. However, the runtime instructions expect the platform to spawn sub-agents and permit committing changes and running arbitrary repo commands. If you enable autonomous invocation for your agent, the skill will exercise those abilities; that is consistent with its purpose but increases the scope of actions the agent can perform on your repository and build environment.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install ralph-mode - 安装完成后，直接呼叫该 Skill 的名称或使用
/ralph-mode触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.2.0
- No file changes detected in this release.
- Documentation, descriptions, and usage instructions remain unchanged from the previous version.
v1.1.0
- No changes detected in this version; documentation and implementation remain the same.
v1.0.0
Version 1.0.0 of ralph-mode introduces a new methodology for autonomous, iterative development loops with structured planning and robust validation gates.
- Implements Ralph Wiggum technique adapted for OpenClaw with autonomous loops and backpressure validation.
- Defines a three-phase workflow: requirements definition, planning, and structured iterative building.
- Enforces programmatic (test, lint, typecheck, build) and subjective (LLM-as-judge) gates for quality assurance.
- Introduces Hats (personas) for specialized agent roles: Architect, Implementer, Tester, Reviewer.
- Provides standard project file structure and operational guides for Next.js, Python, FastAPI, and GPU workloads.
- Includes escape hatches and completion criteria for clear project closure.
元数据
常见问题
Ralph Mode - Autonomous Development Loops 是什么？
Autonomous development loops with iteration, backpressure gates, and completion criteria. Use for sustained coding sessions that require multiple iterations, test validation, and structured progress tracking. Supports Next.js, Python, FastAPI, and GPU workloads with Ralph Wiggum methodology adapted for OpenClaw. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 3428 次。
如何安装 Ralph Mode - Autonomous Development Loops？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install ralph-mode」即可一键安装，无需额外配置。
Ralph Mode - Autonomous Development Loops 是免费的吗？
是的，Ralph Mode - Autonomous Development Loops 完全免费（开源免费），可自由下载、安装和使用。
Ralph Mode - Autonomous Development Loops 支持哪些平台？
Ralph Mode - Autonomous Development Loops 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Ralph Mode - Autonomous Development Loops？
由 richginsberg（@richginsberg）开发并维护，当前版本 v1.2.0。
推荐 Skills