Claude Skill
comet-ml/opik-openclaw
OpenClaw 官方插件,将代理轨迹导出至 Opik。监控代理行为、成本、令牌消耗和错误。专为可观测性和评估设计。
概览
仓库信息
安装这个 Skill
git clone https://github.com/comet-ml/opik-openclaw.gitRegistry 信息
项目简介
OpenClaw 官方插件,用于将代理轨迹导出至 Opik,支持监控代理行为、成本、令牌消耗和错误等信息。
🦞 Official plugin for OpenClaw that exports agent traces to Opik. See and monitor agent behaviour, cost, tokens, errors and more.
要点
- 将 OpenClaw 轨迹无缝导出至 Opik
- 监控代理行为、成本、令牌消耗和错误
- 专为可观测性和评估工作流设计
- 支持 Clawdbot 和 Moltbot 代理
使用场景
- 追踪和调试代理决策轨迹
- 分析每次运行的令牌消耗和成本
- 评估代理随时间变化的性能
- 集成 Opik 实现集中式可观测性
README 摘要
<h1 align="center" style="border-bottom: none"> <div> <a href="https://www.comet.com/site/products/opik/?from=llm&utm_source=opik&utm_medium=github&utm_content=header_img&utm_campaign=opik"> <picture> <source media="(prefers-color-scheme: dark)" srcset="https://raw.githubusercontent.com/comet-ml/opik/refs/heads/main/apps/opik-documentation/documentation/static/img/logo-dark-mode.svg"> <source media="(prefers-color-scheme: light)" srcset="https://raw.githubusercontent.com/comet-ml/opik/refs/heads/main/apps/opik-documentation/documentation/static/img/opik-logo.svg"> <img alt="Comet Opik logo" src="https://raw.githubusercontent.com/comet-ml/opik/refs/heads/main/apps/opik-documentation/documentation/static/img/opik-logo.svg" width="200" /> </picture> </a> <br /> 🔭 OpenClaw Opik Observability Plugin </div> </h1> <p align="center"> Official plugin for <a href="https://github.com/openclaw/openclaw">OpenClaw</a> that exports agent traces to <br/> <a href="https://www.comet.com/docs/opik/">Opik</a> for observability and monitoring. </p> <div align="center"> [](./LICENSE) [](https://www.npmjs.com/package/@opik/opik-openclaw) <img src="screenshot.png" alt="Openclaw on Opik Demo"/> </div> ## Why This Plugin [Opik](https://github.com/comet-ml/opik) is a leading open-source LLM and agent observability, tracing, evaluation and optimization platform. `@opik/opik-openclaw` adds native Opik tracing for OpenClaw runs: - LLM request/response spans - Sub-agent request/response spans - Tool call spans with inputs, outputs, and errors - Run-level finalize metadata - Usage and cost metadat