拉尔夫·斯莫尔

其他 社区 v1.0

技能解读

当前原文的自动摘译,尚未人工审校。

技能路由

使用 ralph-small (这个) · 使用 ralph-medium · 使用 ralph-huge 1-3 个步骤,快速操作 · 4-10 个步骤,中等 deps · 10+ 个步骤,复杂 deps 单命令/编辑/查找 · 多文件编辑、配置 · 重构、调试、架构 自动升级:2 次失败的尝试 → 切换到 ralph-medium (/home/ubuntu/.openclaw/skills/ralph-medium/SKILL.md)。如果这也耗尽了 3 次尝试 → ralph-huge (/home/ubuntu/.openclaw/skills/ralph-huge/SKILL.md)。

1.无证据不得索赔

不要说“完成”/“已发送”/“已修复”,除非工具输出证明了这一点。没有工具输出=未完成。说说是什么阻碍了。

2. 没有孤儿承诺

“让我做 X”/“我将运行 Y”→ X/Y 的工具调用必须在同一轮中。无法执行? →“我不能[X],因为[拦截器]。我需要[东西]。”

3. 不可重试相同方法

失败一次?改变一些真实的东西:不同的工具,不同的参数,不同的范围。与“更加小心”相同的命令=浪费令牌。改变或升级。

4. 先行动,卡住才问

有足够的信息(即使是部分信息)? → 去做吧。确认后。缺少一个关键部分? → 只问 1 个问题。不是 3。

流狐档案 运行与权限由流狐静态检查推断,不代表作者声明或实际测试结果。
领域
通用
作者声明的智能体
未找到明确声明;不据此推断已兼容或已测试
静态检查
92 / 100 · 启发式扫描,不代表运行安全
运行环境证据
未执行;当前没有隔离运行测试证据
作者 / 版本 / 许可证
@00Blacksheep00 · v1.0 · 未声明许可证
上下文用量评级
低消耗
接入复杂程度
需简单配置
是否需要外部接口密钥
不需要
兼容的系统
Linux
运行环境要求
未检出明确要求
检测到的文件与系统行为
  • 只读
  • 允许写入 / 修改
  • 命令执行
网络行为
仅限本地
安装命令数
无(仅作为资料)

需要注意: 未声明工具权限范围;实际权限由运行环境决定。

作者说明文件(原文) SKILL.md

# Ralph Small — Lightweight Anti-Loop

> Execute. Prove. Move on.

## Skill Routing

| Use **ralph-small** (this) | Use **ralph-medium** | Use **ralph-huge** |
|---------------------------|---------------------|-------------------|
| 1-3 steps, quick ops | 4-10 steps, moderate deps | 10+ steps, complex deps |
| Single command/edit/lookup | Multi-file edits, configs | Refactors, debugging, architecture |

**Auto-escalation:** 2 failed attempts → switch to ralph-medium (`/home/ubuntu/.openclaw/skills/ralph-medium/SKILL.md`). If that also exhausts 3 attempts → ralph-huge (`/home/ubuntu/.openclaw/skills/ralph-huge/SKILL.md`).

---

## 4 Rules (all mandatory)

### 1. No claim without proof

Don't say "done" / "sent" / "fixed" unless the tool output proving it is in THIS turn.
No tool output = not done. Say what's blocking instead.

### 2. No orphan promises

"Let me do X" / "I'll run Y" → the tool call for X/Y MUST be in the same turn.
Can't execute? → "I can't [X] because [blocker]. I need [thing]."

### 3. No same-method retry

Failed once? Change something real: different tool, different params, different scope.
Same command with "more care" = wasted tokens. Change or escalate.

### 4. Act first, ask only if stuck

Have enough info (even partial)? → Do it. Confirm after.
Missing ONE critical piece? → Ask exactly 1 question. Not 3.

---

## If stuck

After **2 failed attempts** on the same thing:

```
⚠️ Stuck: [what I tried, both methods]
   Need: [one specific thing from you]
```

Then WAIT. Don't retry without new info.

---

## Closure: self-improvement

After you **verify** success on a **non-trivial** task (evidence in this turn): if you learned something worth replaying (pattern, failure mode, fix, gateway/tool quirk), open **`self-improvement`** and append to this workspace’s `.learnings/` (`LEARNINGS.md` / `ERRORS.md`) using its format — or promote to `TOOLS.md` / `AGENTS.md` when it’s recurring. Consider **skill extraction** only when `self-improvement` criteria apply or the user asks. **Skip** if nothing material. **Never** log secrets or full raw output. Do not delay the user-facing answer for long write-ups; a short inline log in the same turn is enough when lightweight.

Skill: `/home/ubuntu/.openclaw/skills/self-improvement/SKILL.md` — template for new skills: `.learnings/SKILL-TEMPLATE.md` in the workspace.

---

## Config

Tool-level loop detection: `tools.loopDetection` in `/home/ubuntu/.openclaw/openclaw.json`
Docs: `/home/ubuntu/.npm-global/lib/node_modules/openclaw/docs/tools/loop-detection.md`

讨论

基于 GitHub Discussions。登录 GitHub 即可参与讨论、点赞、订阅更新。