← 返回 Skills 市场

LoR (Loop of Resilience)

Name: LoR (Loop of Resilience)
Author: uwe77

作者 Zhang, Yu-Wei · GitHub ↗ · v1.0.1 · MIT-0

cross-platform ✓ 安全检测通过

总下载

当前安装

版本数

在 OpenClaw 中安装

/install lor-skill

功能描述

A modular 5-step engineering verification loop ensuring reliable execution and output validation for complex tasks like coding and planning.

使用说明 (SKILL.md)

SKILL: LoR (Loop of Resilience)

Metadata

Name: LoR (Loop of Resilience)
Category: Engineering, Verification, Chain of Thought (CoT), Agent Reliability
Description: A modular 5-step engineering verification loop for complex tasks. It implements a sophisticated Chain of Thought (CoT) paradigm, automating the Raw Reply -> Verification Planning -> Forced Execution -> Result Checking -> Final Output flow.
Repository: https://github.com/lumi77-mac/lor-skill

Usage

Auto-triggered on complex tasks (coding, architecture, complex planning).
Explicit trigger via /lor [task] or @[AgentName] /lor [task].

Requirements

Node.js (v16.0.0 or higher)
OpenClaw execution environment

Auto-Registration

The system outputs a registration hook [REGISTER_LOR_SKILL] upon installation.
Agent Memory: Agents should detect this hook and register the skill in their MEMORY.md.

Safety Constraints

Require user confirmation for destructive operations (exec, write, edit).
Read-only operations (read, search) execute automatically.
Max 3 retries in Step 4.

安全使用建议

This skill appears internally consistent and limited to local verification logic, but before installing you should: (1) verify you trust the repository owner (setup.sh copies files into ~/.openclaw/skills/lor), (2) confirm your agent enforces the SKILL.md safety constraints (the skill doesn't implement confirmation checks itself), (3) review the [REGISTER_LOR_SKILL] registration behavior your agent will perform (so it doesn't auto-enable the skill without your consent), and (4) run the provided tests locally (npm test / node tests/test_lor.js) to ensure behavior is as expected. If you don't trust the source, do not run the setup script or install the skill.

功能分析

Type: OpenClaw Skill Name: lor-skill Version: 1.0.1 The LoR (Loop of Resilience) skill bundle is a legitimate implementation of a Chain of Thought (CoT) verification framework. The code is transparent and includes a security check in `engine/verifier.js` to prevent path traversal. There are no signs of data exfiltration, malicious execution, or harmful prompt injection; the instructions in `SKILL.md` are strictly aligned with the stated purpose of task verification and agent memory management.

能力评估

✓ Purpose & Capability

Name/description (a 5-step verification/CoT loop) align with the supplied files: modular engine modules, a setup script, README and tests. There are no unexpected binaries, external service credentials, or unrelated capabilities requested.

ℹ Instruction Scope

SKILL.md describes auto-triggering on complex tasks and safety constraints (e.g., require confirmation for destructive ops) but the codebase contains no enforcement for those policies — the skill provides local verification helpers and a verifier that restricts access to the skill directory. The SKILL.md also documents an auto-registration hook ([REGISTER_LOR_SKILL]) which the setup script echoes; actual registration into agent memory depends on the agent implementation. In short: the instructions are descriptive and prescriptive, but much of the enforcement is expected to be handled by the host agent rather than by the skill itself.

ℹ Install Mechanism

There is no remote download/install from arbitrary URLs (low risk). The included scripts/setup.sh copies engine/, SKILL.md and package.json into $HOME/.openclaw/skills/lor — a standard local install behavior. This writes files into the user's home directory, so users should verify the repository origin before running the script. No extracted archives, remote binaries, or network fetches are present.

✓ Credentials

The skill requires only Node.js (engine metadata) and declares no environment variables, secrets, or external credentials. The code uses only core Node modules (fs, path) and the verifier restricts file access to the skill root, which is proportionate to a verification utility.

ℹ Persistence & Privilege

always:false (default) and there are no privileged manifest flags. The setup script prints a registration hook string intended for agents to add to MEMORY.md; this could lead to the skill being persisted by an agent that follows that convention. That persistence depends on agent behavior — the skill itself does not autonomously persist or modify agent state.

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 lor-skill
安装完成后，直接呼叫该 Skill 的名称或使用 /lor-skill 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

v1.0.1

LoR Skill v1.0.1 Changelog - Expanded skill category to include Chain of Thought (CoT). - Updated description to clarify implementation of a Chain of Thought paradigm. - Improved metadata and documentation in SKILL.md for clarity.

v1.0.0

- Initial release of the LoR (Loop of Resilience) skill for engineering verification. - Introduces a modular 5-step loop: Raw Reply, Verification Planning, Forced Execution, Result Checking, and Final Output. - Supports auto-triggering for complex tasks, or manual triggering with a command. - Includes safety constraints: confirmation required for destructive actions, and automatic execution for read-only operations. - Limits result checking step to a maximum of 3 retries. - Provides an auto-registration hook for agent integration.

元数据

Slug lor-skill

版本 1.0.1

许可证 MIT-0

累计安装 0

当前安装数 0

历史版本数 2

常见问题