메뉴
Query Langfuse traces for debugging LLM calls, analyzing token usage, and investigating workflow executions. Use when debugging AI/LLM behavior, checking trace data, or analyzing observability metrics.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | langfuse-traces |
| description | Query Langfuse traces for debugging LLM calls, analyzing token usage, and investigating workflow executions. Use when debugging AI/LLM behavior, checking trace data, or analyzing observability metrics. |
| allowed-tools | Bash, Read |
Query and analyze Langfuse trace data directly from Claude Code.
Invoke the query script using the base directory shown above:
bash <base_directory>/scripts/query.sh <command> [options]
| Command | Args | Description |
|---|---|---|
traces | [limit] [session_id] [name] | List recent traces |
trace | trace_id | Get full trace with observations |
observations | [limit] [trace_id] | List spans/generations |
sessions | [limit] | List sessions |
summary | [limit] | Compact one-line-per-trace view |
# List last 20 traces
bash <base_directory>/scripts/query.sh traces 20
# Get specific trace detail
bash <base_directory>/scripts/query.sh trace tr-abc123
# List observations for a trace
bash <base_directory>/scripts/query.sh observations 50 tr-abc123
# Quick summary of recent activity
bash <base_directory>/scripts/query.sh summary 10
The script reads credentials from .env.local, .env, or environment variables:
LANGFUSE_PUBLIC_KEY=pk-lf-...
LANGFUSE_SECRET_KEY=sk-lf-...
LANGFUSE_BASE_URL=https://cloud.langfuse.com # optional, default
curl (standard on macOS/Linux)jq for JSON parsingAll commands return JSON (piped through jq). Use jq filters for specific fields:
bash <base_directory>/scripts/query.sh traces 5 | jq '.data[].name'
Schedule Slack reminders. Use when the user asks to schedule, remind, or send a Slack message at a specific time.
Send notifications via Telegram. Use when the user asks to notify, alert, message, or ping — or when a long-running task completes and the user previously asked to be notified. Can also schedule messages for later delivery.
Self-messaging capability. Invoke this skill when you need to reframe your approach mid-task, break complex work into deliberate phases, or restart your reasoning with a fresh perspective — without waiting for user input.
Interact with X (Twitter) API v2. Post tweets, search, engage, moderate, and analyze — all from your AI agent. Full 36-command skill for Twitter/X automation.
Manage your Beehiiv newsletter — list and read posts, create drafts from Markdown/HTML files. 6-command skill for Beehiiv newsletter automation.
Generate self-contained HTML broadsheet newspapers to explain any subject worth reporting — project status, reviews, dispatches, reports, or anything else.