メニュー
Troubleshoot ADK errors and issues. Use when encountering API errors, model overloaded errors, state issues, tool failures, authentication problems, or unexpected agent behavior.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
| name | adk-debugger |
| description | Troubleshoot ADK errors and issues. Use when encountering API errors, model overloaded errors, state issues, tool failures, authentication problems, or unexpected agent behavior. |
| Symptom | Likely Cause | Quick Fix |
|---|---|---|
GOOGLE_API_KEY not set | Missing .env | Create app/.env |
503 model overloaded | Gemini 3 rate limit | Switch to Gemini 2.5 Pro |
| Tool not being called | output_schema set | Remove output_schema |
| State variable empty | Key mismatch | Check output_key names |
| TTS fails | Vertex AI mode | Use AI Studio for multi-speaker |
Error: GOOGLE_API_KEY not set or Invalid API key
Fix: Create .env file in app/ folder (not project root):
echo "GOOGLE_API_KEY=your_key" >> app/.env
echo "MAPS_API_KEY=your_maps_key" >> app/.env
Error: 503 UNAVAILABLE - model overloaded
Fix: Edit app/config.py:
# Switch from Gemini 3 to 2.5
FAST_MODEL = "gemini-2.5-pro" # More stable
PRO_MODEL = "gemini-2.5-pro"
Cause: Using output_schema disables tool calling.
Fix: Remove output_schema or use separate agent:
# DON'T: This disables tools
agent = LlmAgent(
tools=[my_tool],
output_schema=MySchema, # Conflicts!
)
# DO: Separate concerns
tool_agent = LlmAgent(tools=[my_tool], ...)
schema_agent = LlmAgent(output_schema=MySchema, ...)
Cause: output_key doesn't match instruction placeholder.
Fix: Ensure names match exactly:
# Agent 1
agent1 = LlmAgent(output_key="market_data", ...)
# Agent 2 instruction must use same key
INSTRUCTION = "Use this data: {market_data}" # Must match!
cat app/.envmake test-intake[See references/common-errors.md for complete error catalog]
Tests generated ADK agent code through a save-first validation pipeline. Saves code to disk FIRST, then tests the saved code via start_agent and talk_to_agent. Drives the iterative refinement loop.
FastAPI development best practices and conventions. Use when building, reviewing, or debugging FastAPI applications, REST APIs, or Pydantic models.
Reviews Python code for quality, style, and common bugs. Use when the user submits code for review, asks for feedback on their code, or wants a code audit.
Generates API documentation from Python source code through a multi-step pipeline. Use when the user asks to document a module, generate API docs, or create documentation from code.
Plans a new software project by gathering requirements through structured questions before producing a plan. Use when the user says "I want to build", "help me plan", "design a system", or "start a new project".
Generates structured technical reports in Markdown. Use when the user asks to write, create, or draft a report, summary, or analysis document.