← 返回 Skills 市场

Auto Skill 20260404 063632

Name: Auto Skill 20260404 063632
Author: timo2026

作者 Timo2026 · GitHub ↗ · v1.0.0 · MIT-0

cross-platform ⚠ suspicious

总下载

当前安装

版本数

在 OpenClaw 中安装

/install auto-skill-20260404-063632

功能描述

API问题对抗讨论求解器 - 4模型线性对抗找出最优解

使用说明 (SKILL.md)

API问题对抗讨论求解器

概述

当遇到复杂的API问题时，使用4模型线性对抗找出最优解：

架构师：问题诊断+方案设计
审核者：质疑风险+建议验证
仲裁者：综合判断+最终决策
收敛者：找共同点+整理输出

使用场景

API调用失败需要诊断
多个解决方案需要选择最优
需要多个角度分析问题

本次应用案例

问题描述

API Key验证失败，请求超时

诊断过程

Step 1: 网络诊断

ping测试：208ms延迟，3包全收 ✅
HTTPS请求：超时（30秒无响应）❌

Step 2: 配置检查

base_url: coding.dashscope.aliyuncs.com/v1 ✅
API Key格式: sk-sp-xxxxx ✅

Step 3: 降级方案

当前会话fallback模式 ✅

最优解

网络诊断 → 有限重试 → 模拟降级

代码实现

线性对抗引擎

# 4模型线性排列
models = [
    {"角色": "架构师", "延迟": 5},
    {"角色": "审核者", "延迟": 5},
    {"角色": "仲裁者", "延迟": 8},
    {"角色": "收敛者", "延迟": 0},
]

# 依次发言，记录回复，最终收敛
for model in models:
    response = call_model(model)
    save_cache(response)
    time.sleep(model["延迟"])

教训总结

base_url必须匹配API类型（Coding Plan用专用URL）
网络层正常不代表应用层正常
需要降级机制保证可用性

相关文件

modules/linear_adversarial_4model.py - 线性对抗引擎
tests/test_basic.py - 单元测试
memory/api_debate_analysis_20260404.txt - 对抗分析报告

生成时间: 2026-04-04 作者: 海狸 🦫

安全使用建议

This skill appears to describe a plausible debugging workflow, but there are unexplained gaps: SKILL.md mentions modules and memory files that are not included, and it references a specific API endpoint and API-key format while declaring no required credentials. Before installing or using it: (1) ask the publisher for the missing code files or an explanation of where call_model/save_cache operate; (2) do not provide API keys or secrets until you confirm how and where they will be used/stored; (3) prefer running the workflow in a sandboxed environment; (4) request a homepage or source repo to inspect actual implementation; and (5) if you need to trust this skill, insist that it explicitly declare required environment variables and storage locations so you can assess credential exposure.

能力评估

ℹ Purpose & Capability

The skill claims to perform a 4-model linear adversarial analysis for API debugging — the prose and pseudocode in SKILL.md are consistent with that purpose. However, it references an explicit base_url (coding.dashscope.aliyuncs.com/v1) and an API key format (sk-sp-xxxxx) that imply integration with an external API/service even though the skill declares no required environment variables or credentials. This mismatch is unexplained.

ℹ Instruction Scope

Instructions include logical runtime steps (ping, HTTPS request, configuration checks, invoking call_model, save_cache). They do not explicitly instruct the agent to read arbitrary system files or exfiltrate data, but they reference saving cache and named files (modules/linear_adversarial_4model.py, memory/api_debate_analysis_20260404.txt) that are not present in the bundle. It's unclear where saved data is written or what 'call_model' targets — this gives the agent discretionary I/O/network behavior that isn't fully specified.

✓ Install Mechanism

This is an instruction-only skill with no install spec and no code files included, so nothing is written to disk or installed by the skill itself. That minimizes installation risk.

ℹ Credentials

No environment variables or credentials are declared, yet the documentation mentions a concrete base_url and an API key pattern. If the skill is intended to call an external API, the absence of declared required credentials is a proportionality/information mismatch: users might be prompted later to supply secrets or store them in unspecified locations. The skill does not ask for broad system credentials, but the missing explicit credential handling is a concern.

✓ Persistence & Privilege

always is false and there are no special persistence or cross-skill configuration changes declared. The skill does reference saving cache and a memory file name, but those are internal references in the docs; the package does not request persistent privileges or system-wide changes.

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 auto-skill-20260404-063632
安装完成后，直接呼叫该 Skill 的名称或使用 /auto-skill-20260404-063632 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

v1.0.0

- Initial release of api-debate-solver (v1.0.0). - Implements 4-model linear adversarial process for API problem solving. - Supports diagnosis and solution optimization for API failures, including network and configuration checks. - Includes fallback mode for degraded operation. - Provides sample code for linear process orchestration. - Adds documentation and basic unit tests.

元数据

Slug auto-skill-20260404-063632

版本 1.0.0

许可证 MIT-0

累计安装 0

当前安装数 0

历史版本数 1

常见问题