인터페이스는 현지화되어 있으며, Skill 소개와 설명 문서는 선택한 언어로 AI가 자동 번역합니다. 코드·명령어·경로·전문 용어는 원문을 유지하며 원작자 문서가 우선합니다.
Skills 둘러보기
생산성 및 협업 · SKILL.md

sdk-and-tracing

sdk-and-tracing

Traceloop SDK initialization, decorator tracing, manual spans, and client-facing observability workflows.

Independent GitHub snapshot · Automated classification · Original documentation governs

Skill 문서

GitHub

name: sdk-and-tracing description: "Traceloop SDK initialization, decorator tracing, manual spans, and client-facing observability workflows." disable-model-invocation: true metadata: disco-role: operating license: Apache 2.0

sdk-and-tracing

Use this sub-skill for Traceloop SDK bootstrap, offline tracing, and user-app tracing workflows.

Use it when

  • you need to initialize Traceloop
  • you need exporter or processor setup for local debugging or Traceloop/OTLP delivery
  • you need decorator-based spans from @workflow, @task, @agent, @tool, or @conversation
  • you need manual LLM spans via track_llm_call
  • you need association properties or prompt context on spans
  • you need the prompt, dataset, experiment, user-feedback, or guardrail surfaces at a high level

Install choices

  • Base SDK: traceloop-sdk
  • Dataset helpers: traceloop-sdk[datasets] when you need CSV/DataFrame helpers
  • Debug exporters: InMemorySpanExporter for assertions, ConsoleSpanExporter for inspection

Route elsewhere

Read first

Runtime helpers

Operating notes

  • use_attributes=True is the current default; use_legacy_attributes is deprecated.
  • TRACELOOP_TRACE_CONTENT=false intentionally hides decorated entity inputs/outputs and prompt content.
  • If Traceloop.init() receives both exporter and processor, the exporter is ignored; wrap the exporter inside the processor instead.
  • Traceloop.init() returns a client only when it is using the Traceloop endpoint path with an API key and no custom exporter or processor.
  • Stay at SDK/bootstrap depth here; do not descend into provider wrapper internals or exact semantic-convention tables.
저장한 Skills