← 返回 Skills 市场

Skill

Name: Skill
Author: amitbidlan

作者 Amit Bidlan · GitHub ↗ · v0.1.0 · 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 lumin

功能描述

Observe and debug your OpenClaw agent with Lumin — local-first AI observability. Records every LLM call, tool call, and decision; view at localhost:3000. Use...

使用说明 (SKILL.md)

Lumin — Local Agent Observability

Lumin records every LLM call, tool call, and decision your OpenClaw agent makes. View them at \x3Chttp://localhost:3000>. Data never leaves the machine — everything runs in a single Docker container.

This skill provides three commands that hit Lumin's local HTTP API so you can check status and pull trace data without leaving the chat.

Check if Lumin is running

Run: node $SKILL_DIR/lumin.mjs status

Returns one of:

Lumin is running on http://localhost:8000
Lumin is not running on …. Start with: docker run -p 3000:3000 -p 8000:8000 zistica/lumin

Show recent traces

Run: node $SKILL_DIR/lumin.mjs traces

Returns the last 5 agent traces formatted like:

Recent agent traces:
  1. openclaw_session  3.2s  $0.0023  OK
  2. openclaw_session  1.8s  $0.0011  OK
  3. openclaw_session  5.1s  $0.0089  ERROR
  …
Open http://localhost:3000/traces for the full list.

Show one trace's spans

Run: node $SKILL_DIR/lumin.mjs trace \x3Cid>

Returns the trace metadata plus every span with type, model, tokens, cost, and any error message.

Wire OpenClaw to Lumin

This skill only reads from Lumin. To get your OpenClaw agent's traces into Lumin, install the companion exporter:

npm install @lumin-io/openclaw

…and pass luminProcessor() into your OpenClaw OTel config. See \x3Chttps://github.com/amitbidlan/zistica-lumin/tree/main/packages/integrations/openclaw>.

Start Lumin (if not running)

docker run -p 3000:3000 -p 8000:8000 zistica/lumin

Then open \x3Chttp://localhost:3000>.

安全使用建议

This skill appears safe for its stated purpose if you understand that agent traces can contain sensitive details. Prefer the default localhost setup for local-only use, verify the companion npm package and Docker image before installing them, and only configure hosted Lumin credentials for a service you trust.

功能分析

Type: OpenClaw Skill Name: lumin Version: 0.1.0 The 'lumin' skill is a local observability tool for OpenClaw agents that retrieves and displays execution traces. The core logic in `lumin.mjs` uses the native Node.js fetch API to query a local or user-defined endpoint (defaulting to localhost:8000) for health status and trace metadata. The code is well-structured, includes proper URL encoding for parameters, implements timeouts, and lacks any indicators of data exfiltration, malicious execution, or prompt injection.

能力标签

requires-oauth-tokenrequires-sensitive-credentials

能力评估

ℹ Purpose & Capability

The stated observability purpose matches the helper code, which only checks Lumin status and reads recent traces/spans, but those traces can describe LLM calls, tool calls, decisions, costs, tokens, sessions, and errors.

✓ Instruction Scope

The documented commands are user-directed status/read operations; the helper does not mutate traces, delete data, install software automatically, or run hidden background work.

ℹ Install Mechanism

The skill itself has no install spec, but the documentation recommends a companion npm package and Docker image without pinned versions; this is purpose-aligned but should be verified before use.

ℹ Credentials

The default endpoint is localhost, but the artifacts also support a configurable LUMIN_HOST and mention an optional hosted-instance bearer token, so the local-only privacy claim should be treated as conditional.

ℹ Persistence & Privilege

The helper does not request elevated OS privileges, but Lumin’s intended function is to retain agent observability traces, which may include sensitive run history.

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

v0.1.0

Initial release of the Lumin skill for local agent observability. - Observe and debug OpenClaw agents with Lumin via three chat commands: check status, show recent traces, and inspect a specific trace. - Records every LLM call, tool call, and agent decision; view them at http://localhost:3000. - All data is local—no external servers involved. - Connects to Lumin's API using configurable environment variables (LUMIN_HOST, LUMIN_API_KEY). - To export traces from OpenClaw, use the companion npm package: @lumin-io/openclaw. - Quick start with Docker; detailed instructions provided.

元数据

Slug lumin

版本 0.1.0

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

Skill 是什么？

Observe and debug your OpenClaw agent with Lumin — local-first AI observability. Records every LLM call, tool call, and decision; view at localhost:3000. Use... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 31 次。

如何安装 Skill？

在 OpenClaw 或 Claude Code 对话框中运行命令「/install lumin」即可一键安装，无需额外配置。

Skill 是免费的吗？

是的，Skill 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。

Skill 支持哪些平台？

Skill 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 Skill？

由 Amit Bidlan（@amitbidlan）开发并维护，当前版本 v0.1.0。