메뉴
Use when adding neatlogs observability to a Python project that uses the Google Agent Development Kit (imports `google.adk`, builds an ADK agent/runner).
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
Use when adding neatlogs observability to a Python project that uses LangChain or LangGraph (imports `langchain*` / `langgraph`, builds chains, runnables, or a graph).
Use when adding neatlogs observability to a TypeScript/Node.js project that uses LangChain or LangGraph (depends on `@langchain/*` / `@langchain/langgraph`).
Use when adding neatlogs observability to a Python project that uses CrewAI (imports `crewai`, builds a Crew/Flow with agents and tasks).
Use when adding neatlogs observability to a Go project — Google Gemini (genai), Google ADK agents, or custom code. Covers Init, WrapGenAI, ADK passthrough, Trace, and Identify (sessions & end-users).
Use when adding neatlogs observability to a Python project that uses the Agno agent framework (imports `agno`, constructs Agno `Agent`s).
Use when adding neatlogs observability to a Python project that uses DSPy (imports `dspy`, defines `dspy.Module`s / signatures).
SOC 직업 분류 기준
| name | neatlogs-py-google-adk |
| description | Use when adding neatlogs observability to a Python project that uses the Google Agent Development Kit (imports `google.adk`, builds an ADK agent/runner). |
| compatibility | Neatlogs Wizard Agent |
| metadata | {"author":"neatlogs","version":"1.0","language":"python","framework":"google-adk"} |
This project uses the Google ADK (google.adk.runners.Runner / InMemoryRunner, LlmAgent). Neatlogs instruments it with neatlogs.wrap(runner).
neatlogs.wrap(runner)neatlogs.wrap() patches the Runner's run() (sync generator) and run_async() (async generator), tracing each invocation as a WORKFLOW span with token usage, output, and tool-call metadata extracted from the event stream.
WORKFLOW google_adk.runner.run_async (per invocation)
Combine with @neatlogs.span / neatlogs.trace / neatlogs.log for your own orchestration; the runner span nests under them.
references/1-install.mdreferences/2-add-init.mdreferences/3-set-env.mdreferences/4-wrap-runner.mdreferences/5-spans-trace-log.mdreferences/6-flush-shutdown.mdneatlogs.init() MUST run BEFORE importing google.adk. load_dotenv() runs before init().runner = neatlogs.wrap(InMemoryRunner(agent=..., app_name=...)). Returns the same instance; run_async() stays an async generator (consume with async for).runner.run_async() loop in @span/trace — wrap() already opens the WORKFLOW span. Use @span for YOUR orchestration only.os.getenv(). ADK reads GOOGLE_API_KEY (set GOOGLE_GENAI_USE_VERTEXAI=0 for AI-Studio keys).import neatlogs at module top level.references/5-spans-trace-log.mdreferences/span-kinds.mdidentify()) → references/sessions-and-end-users.md