encounter/adk-python
0
0
Fork 0
mirror of https://github.com/encounter/adk-python.git synced 2026-07-09 18:19:28 -07:00
Code Issues Packages Projects Releases Wiki Activity

docs: Update agent builder instructions for clarity and best practices for ADK specific requirements

Browse Source 
PiperOrigin-RevId: 822862453
This commit is contained in:
George Weale
2025-10-22 21:31:40 -07:00
committed by Copybara-Service
parent 055e86a9d4
commit 9cf7ab11a8
1 changed files with 6 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
contributing/samples/adk_agent_builder_assistant/instruction_embedded.template
+6 -3
View File
@@ -393,11 +393,13 @@ def log_tool_result(tool: BaseTool, tool_args: Dict[str, Any], tool_context: Too
**File Naming & Structure:**
- Main configuration MUST be `root_agent.yaml` (not `agent.yaml`)
- Agent directories need `__init__.py` with `from . import agent`
- **Tools directory MUST have `__init__.py`** - The `tools/` folder requires an empty `__init__.py` file to be a valid Python package (required for imports)
- Place each tool in the `tools/` package using one module per tool (for example, `tools/dice_tool.py`).
Add an empty `tools/__init__.py` so imports such as `project_name.tools.dice_tool.roll_dice` work.
- Python files in agent directory, YAML at root level
**Tool Configuration:**
- Function tools: `project_name.tools.module.function_name` format (all dots, must start with project folder name)
- Function tools: Use dotted import paths that start with the project folder name
(e.g., `project_name.tools.dice_tool.roll_dice`)
- No `.py` extension in tool paths
- No function declarations needed in YAML
- **Critical**: Tool paths must include the project folder name as the first component (final component of project folder path only)
@@ -407,7 +409,7 @@ def log_tool_result(tool: BaseTool, tool_args: Dict[str, Any], tool_context: Too
- **SequentialAgent**: NO `model` field - workflow agent that orchestrates other agents in sequence
- **ParallelAgent**: NO `model` field - workflow agent that runs multiple agents in parallel
- **LoopAgent**: NO `model` field - workflow agent that executes agents in a loop
- **CRITICAL**: Only LlmAgent accepts a model field. Workflow agents (Sequential/Parallel/Loop) do NOT have model fields
- **CRITICAL**: Only LlmAgent accepts a model field. Workflow agents (Sequential/Parallel/Loop) do NOT have model fields or tool lists; they orchestrate `sub_agents` that provide tooling.
**ADK AgentConfig Schema Compliance:**
- Always reference the embedded ADK AgentConfig schema to verify field requirements
@@ -447,6 +449,7 @@ def log_tool_result(tool: BaseTool, tool_args: Dict[str, Any], tool_context: Too
2. No redundant suggest_file_path calls for pre-approved paths
3. Generated configurations pass schema validation (automatically checked)
4. Follow ADK naming and organizational conventions
5. Every agent configuration explicitly sets `agent_class` and the value matches the agent role; custom classes use a fully qualified dotted path
6. Include clear, actionable instructions for each agent
7. Use appropriate tools for intended functionality