← 返回 Skills 市场
116
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install nmvnb
功能描述
Manages free AI models from OpenRouter for OpenClaw. Automatically ranks models by quality, configures fallbacks for rate-limit handling, and updates opencla...
安全使用建议
This skill appears to do what it says (manage free OpenRouter models and update OpenClaw config), but there are metadata inconsistencies you should resolve before installing:
- It requires an OPENROUTER_API_KEY (SKILL.md and skill.json) despite the top-level registry data claiming no required env vars. Expect to provide a key and consider creating a dedicated OpenRouter key for this use. Do not reuse high-privilege keys.
- The package includes setup.py and console entry points; installation runs pip install -e . in the skill directory. That will install local Python code into your environment and add CLI commands. Review the included files (main.py, watcher.py) yourself (they are present and readable) before installing.
- The skill will modify ~/.openclaw/openclaw.json (agents.defaults.model, fallbacks, models) and create cache/state files under ~/.openclaw. Back up your openclaw.json before running freeride auto or the watcher.
- Network calls are only to openrouter.ai endpoints; no hidden remote endpoints were found. If you want to be cautious, run the tool in a restricted environment first or inspect/modify the code (e.g., remove auto-daemon behavior) and run a dry-run.
If you are comfortable with the author/source, proceed after backing up config and using a dedicated OpenRouter API key. If the source is unknown or untrusted, ask the publisher to clarify the registry metadata mismatch and provide a verified upstream repository before installing.
功能分析
Type: OpenClaw Skill
Name: nmvnb
Version: 1.0.0
The FreeRide skill bundle is a legitimate utility designed to manage and rotate free AI models from OpenRouter within the OpenClaw environment. The code in `main.py` and `watcher.py` performs transparent configuration updates to `~/.openclaw/openclaw.json` and interacts only with the official OpenRouter API to fetch model metadata and test availability. No evidence of data exfiltration, obfuscation, or malicious intent was found; the instructions in `SKILL.md` are consistent with the tool's stated purpose of automating model switching to avoid rate limits.
能力评估
Purpose & Capability
The skill is designed to find/rank free OpenRouter models and update OpenClaw config; the code reads OpenRouter APIs and writes ~/.openclaw/openclaw.json which is coherent with the stated purpose. However, registry-level metadata provided above (no required env vars / no primary credential) contradicts the bundled skill.json and SKILL.md which clearly require OPENROUTER_API_KEY.
Instruction Scope
SKILL.md instructs the agent to check OPENROUTER_API_KEY, run the included CLI (freeride/freeride-watcher), and restart the OpenClaw gateway. The runtime instructions and code only read/write OpenClaw config and call OpenRouter endpoints; they do not attempt to read unrelated system files or exfiltrate data to unknown endpoints.
Install Mechanism
The registry metadata here said 'no install spec', but the package includes setup.py, entry points, and skill.json contains an install command (npx clawhub... && pip install -e .). The install is local (pip install -e .) and only depends on 'requests' — no obscure remote binaries or shortened/unknown URLs. The mismatch between 'no install spec' and the included install instructions is an inconsistency to check.
Credentials
Although the top-level metadata listed no required env vars, the SKILL.md, skill.json and code require OPENROUTER_API_KEY and will read it from the environment or from ~/.openclaw/openclaw.json. Requesting access to the user's OpenRouter API key is reasonable for this function, but the metadata mismatch (missing required env at registry level) is an incoherence and should be resolved before trusting the skill.
Persistence & Privilege
always:false and default autonomous invocation are present (normal). The skill writes cache/state under ~/.openclaw (e.g., .freeride-cache.json, .freeride-watcher-state.json) and updates openclaw.json keys only in agents.defaults.model and agents.defaults.models as described. It does not request system-wide privileges or modify other skills' configs.
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 nmvnb - 安装完成后，直接呼叫该 Skill 的名称或使用
/nmvnb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 with new free AI model management for OpenClaw via OpenRouter:
- Introduces the FreeRide skill to automate selection, ranking, and fallback setup of free AI models.
- Adds CLI tool (`freeride`) and optional watcher utility for auto-rotation on rate limits.
- Updates OpenClaw configuration with optimal free model and dynamic fallback chain—preserves user’s other settings.
- Provides detailed usage instructions, troubleshooting steps, and command reference.
- Removes legacy molt.py script; all functionality migrated to new components.
元数据
常见问题
SEEK 是什么？
Manages free AI models from OpenRouter for OpenClaw. Automatically ranks models by quality, configures fallbacks for rate-limit handling, and updates opencla... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 116 次。
如何安装 SEEK？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install nmvnb」即可一键安装，无需额外配置。
SEEK 是免费的吗？
是的，SEEK 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
SEEK 支持哪些平台？
SEEK 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 SEEK？
由 pz33y（@pz33y）开发并维护，当前版本 v1.0.0。
推荐 Skills