Name: agno_agent
Author: zzw4257

Prompt file imported from zzw4257/free-to-do-app (.cursor/commands/agno_agent.md). Copyright stays with the author.

Agno Agent Development Quick Commands

Overview

This guide covers development of Agno Agent Tools - the AI-powered todo management toolkit based on the Agno framework.

The FreeTodoToolkit provides a set of tools for the Agno Agent to manage todos. For the complete list of available tools, please refer to the source code in llm/agno_tools/tools/ directory.

🏗️ Architecture

Directory Structure

lifetrace/
├── config/prompts/agno_tools/     # Localized messages & prompts
│   ├── zh/                        # Chinese messages
│   └── en/                        # English messages (same structure)
│
├── llm/agno_tools/                # Python implementation
│   ├── __init__.py                # Module exports
│   ├── base.py                    # Message loader (AgnoToolsMessageLoader)
│   ├── toolkit.py                 # Main FreeTodoToolkit class
│   └── tools/                     # Individual tool implementations (organized by category)
│
└── observability/                 # Agent monitoring (Phoenix + OpenInference)
    ├── __init__.py                # Module exports
    ├── config.py                  # Observability configuration
    ├── setup.py                   # Initialization entry point
    └── exporters/
        ├── __init__.py
        └── file_exporter.py       # Local JSON file exporter

Design Patterns

Mixin Pattern: Each tool category is a separate mixin class
Composition: FreeTodoToolkit inherits from all mixins + Agno Toolkit
i18n: Messages loaded from language-specific YAML files
Lazy Loading: Database and LLM clients initialized on-demand

🔧 Adding a New Tool

Step 1: Add Messages (Both Languages)

Create or update YAML files in config/prompts/agno_tools/zh/ and en/:

# config/prompts/agno_tools/zh/my_tool.yaml
my_tool_success: "操作成功: {result}"
my_tool_failed: "操作失败: {error}"
my_tool_prompt: |
  这是给 LLM 的提示词模板。
  参数: {param}

# config/prompts/agno_tools/en/my_tool.yaml
my_tool_success: "Operation successful: {result}"
my_tool_failed: "Operation failed: {error}"
my_tool_prompt: |
  This is a prompt template for LLM.
  Parameter: {param}

Step 2: Create Tool Mixin

Create a new file in llm/agno_tools/tools/:

# llm/agno_tools/tools/my_tools.py
"""My Tools - Description of what these tools do."""

from __future__ import annotations
from typing import TYPE_CHECKING

from lifetrace.llm.agno_tools.base import get_message
from lifetrace.util.logging_config import get_logger

if TYPE_CHECKING:
    from lifetrace.repositories.sql_todo_repository import SqlTodoRepository

logger = get_logger()


class MyTools:
    """My tools mixin"""

    lang: str
    todo_repo: "SqlTodoRepository"  # If needed

    def _msg(self, key: str, **kwargs) -> str:
        return get_message(self.lang, key, **kwargs)

    def my_tool_method(self, param: str) -> str:
        """Tool description for LLM to understand when to use it

        Args:
            param: Description of the parameter

        Returns:
            Result message
        """
        try:
            # Implementation
            result = f"processed {param}"
            return self._msg("my_tool_success", result=result)
        except Exception as e:
            logger.error(f"Failed: {e}")
            return self._msg("my_tool_failed", error=str(e))

Step 3: Register in Toolkit

Update llm/agno_tools/tools/__init__.py:

from lifetrace.llm.agno_tools.tools.my_tools import MyTools

__all__ = [..., "MyTools"]

Update llm/agno_tools/toolkit.py:

from lifetrace.llm.agno_tools.tools import (
    ...,
    MyTools,
)

class FreeTodoToolkit(
    ...,
    MyTools,  # Add mixin
    Toolkit,
):
    def __init__(self, lang: str = "en", **kwargs):
        ...
        tools = [
            ...,
            self.my_tool_method,  # Register tool
        ]

📝 Message Configuration

YAML Structure

Messages are organized by functionality in config/prompts/agno_tools/{lang}/ directory. Each YAML file corresponds to a category of messages.

Message Format

Use {placeholder} for variable substitution
Multi-line prompts use YAML | syntax
Keep messages concise and informative

# Simple message with placeholder
create_success: "Created todo #{id}: {name}"

# Multi-line prompt
breakdown_prompt: |
  Break down this task into subtasks.

  Task: {task_description}

  Return JSON format.

Accessing Messages

# In tool methods
def _msg(self, key: str, **kwargs) -> str:
    return get_message(self.lang, key, **kwargs)

# Usage
return self._msg("create_success", id=123, name="Buy groceries")

🌐 Internationalization

Language Selection

Language is passed through the call chain:

Request Header (Accept-Language)
    ↓
Chat Router (get_request_language)
    ↓
AgnoAgentService(lang=lang)
    ↓
FreeTodoToolkit(lang=lang)
    ↓
AgnoToolsMessageLoader(lang)

Adding a New Language

Create new directory: config/prompts/agno_tools/{lang}/
Copy all YAML files from en/
Translate all messages
The loader will automatically detect the new language

🧪 Testing Tools

Quick Test Script

from lifetrace.llm.agno_tools import FreeTodoToolkit

# Test Chinese
toolkit_zh = FreeTodoToolkit(lang="zh")
print(toolkit_zh.list_todos(status="active", limit=5))

# Test English
toolkit_en = FreeTodoToolkit(lang="en")
print(toolkit_en.list_todos(status="active", limit=5))

Running Tests

uv run python -c "
from lifetrace.llm.agno_tools import FreeTodoToolkit
tk = FreeTodoToolkit(lang='zh')
print(tk.parse_time('明天下午3点'))
"

📊 Observability (Agent Monitoring)

The Agno Agent integrates with Arize Phoenix + OpenInference for tracing and monitoring.

Features

Local JSON Export: Cursor-friendly trace files for AI analysis
Phoenix UI: Optional web-based visualization
Minimal Terminal Output: One-line summary per trace

Configuration

In config/config.yaml:

observability:
  enabled: true                    # Enable observability
  mode: both                       # local | phoenix | both
  local:
    traces_dir: traces/            # Trace file directory
    max_files: 100                 # Max files to keep
    pretty_print: true             # Format JSON for readability
  phoenix:
    endpoint: http://localhost:6006
    project_name: freetodo-agent
  terminal:
    summary_only: true             # One-line output (recommended)

Trace File Format

Each agent run generates a JSON file in data/traces/:

{
  "trace_id": "e078e147372a",
  "timestamp": "2026-01-23T08:23:48.377470+00:00",
  "duration_ms": 26910.94,
  "agent": "breakdown_task",
  "input": "{\"task_description\": \"Make a video\"}",
  "output_preview": "Task breakdown:\n1. Define topic...",
  "tool_calls": [
    {
      "name": "breakdown_task",
      "args": {"task_description": "Make a video"},
      "result_preview": "Task breakdown...",
      "duration_ms": 26910.94
    }
  ],
  "llm_calls": [],
  "status": "success",
  "span_count": 1
}

Terminal Output

With summary_only: true:

[Trace] e078e147372a | 1 tools | 26.91s | traces/20260123_082348_e078e147372a.json

Using Phoenix UI (Optional)

# Start Phoenix server
uv run phoenix serve

# Access http://localhost:6006

✅ Development Checklist

When adding new tools:

Create YAML messages in both zh/ and en/ directories
Create tool mixin class with proper type hints
Add docstrings for LLM to understand tool usage
Use _msg() for all user-facing messages
Handle exceptions and return error messages
Register tool in tools/__init__.py
Add mixin to FreeTodoToolkit class
Register method in tools list
Test with both languages

Install agno_agent and it becomes one more candidate the router can pick - when it fits.

A listing is a routing target with a manifest: what it does, which domains it covers, what it costs and who publishes it. Once installed it sits beside your own models and tools, is scored like any other candidate for each request, and shows up in the trace when it wins. Ratings come from workspaces that installed it, one per account.

# Agno Agent Development Quick Commands ## Overview This guide covers development of **Agno Agent Tools** - the AI-powered todo management toolkit based on the [Agno framework](https://docs.agno.com/). The FreeTodoToolkit provides a set of tools for the Agno Agent to manage todos. For the complete list of available tools, please refer to the source code in `llm/agno_tools/tools/` directory. --- ## 🏗️ Architecture ### Directory Structure ``` lifetrace/ ├── config/prompts/agno_tools/ # Localized messages & prompts │ ├── zh/ # Chinese messages │ └── en/ # English messages (same structure) │ ├── llm/agno_tools/ # Python implementation │ ├── __init__.py # Module exports │ ├── base.py # Message loader (AgnoToolsMessageLoader) │ ├── toolkit.py # Main FreeTodoToolkit class │ └── tools/ # Individual tool implementations (organized by category) │ └── observability/ # Agent monitoring (Phoenix + OpenInference) ├── __init__.py # Module exports ├── config.py # Observability configuration ├── setup.py # Initialization entry point └── exporters/ ├── __init__.py └── file_exporter.py # Local JSON file exporter ``` ### Design Patterns - **Mixin Pattern**: Each tool category is a separate mixin class - **Composition**: FreeTodoToolkit inherits from all mixins + Agno Toolkit - **i18n**: Messages loaded from language-specific YAML files - **Lazy Loading**: Database and LLM clients initialized on-demand --- ## 🔧 Adding a New Tool ### Step 1: Add Messages (Both Languages) Create or update YAML files in `config/prompts/agno_tools/zh/` and `en/`: ```yaml # config/prompts/agno_tools/zh/my_tool.yaml my_tool_success: "操作成功: {result}" my_tool_failed: "操作失败: {error}" my_tool_prompt: | 这是给 LLM 的提示词模板。参数: {param} ``` ```yaml # config/prompts/agno_tools/en/my_tool.yaml my_tool_success: "Operation successful: {result}" my_tool_failed: "Operation failed: {error}" my_tool_prompt: | This is a prompt template for LLM. Parameter: {param} ``` ### Step 2: Create Tool Mixin Create a new file in `llm/agno_tools/tools/`: ```python # llm/agno_tools/tools/my_tools.py """My Tools - Description of what these tools do.""" from __future__ import annotations from typing import TYPE_CHECKING from lifetrace.llm.agno_tools.base import get_message from lifetrace.util.logging_config import get_logger if TYPE_CHECKING: from lifetrace.repositories.sql_todo_repository import SqlTodoRepository logger = get_logger() class MyTools: """My tools mixin""" lang: str todo_repo: "SqlTodoRepository" # If needed def _msg(self, key: str, **kwargs) -> str: return get_message(self.lang, key, **kwargs) def my_tool_method(self, param: str) -> str: """Tool description for LLM to understand when to use it Args: param: Description of the parameter Returns: Result message """ try: # Implementation result = f"processed {param}" return self._msg("my_tool_success", result=result) except Exception as e: logger.error(f"Failed: {e}") return self._msg("my_tool_failed", error=str(e)) ``` ### Step 3: Register in Toolkit Update `llm/agno_tools/tools/__init__.py`: ```python from lifetrace.llm.agno_tools.tools.my_tools import MyTools __all__ = [..., "MyTools"] ``` Update `llm/agno_tools/toolkit.py`: ```python from lifetrace.llm.agno_tools.tools import ( ..., MyTools, ) class FreeTodoToolkit( ..., MyTools, # Add mixin Toolkit, ): def __init__(self, lang: str = "en", **kwargs): ... tools = [ ..., self.my_tool_method, # Register tool ] ``` --- ## 📝 Message Configuration ### YAML Structure Messages are organized by functionality in `config/prompts/agno_tools/{lang}/` directory. Each YAML file corresponds to a category of messages. ### Message Format - Use `{placeholder}` for variable substitution - Multi-line prompts use YAML `|` syntax - Keep messages concise and informative ```yaml # Simple message with placeholder create_success: "Created todo #{id}: {name}" # Multi-line prompt breakdown_prompt: | Break down this task into subtasks. Task: {task_description} Return JSON format. ``` ### Accessing Messages ```python # In tool methods def _msg(self, key: str, **kwargs) -> str: return get_message(self.lang, key, **kwargs) # Usage return self._msg("create_success", id=123, name="Buy groceries") ``` --- ## 🌐 Internationalization ### Language Selection Language is passed through the call chain: ``` Request Header (Accept-Language) ↓ Chat Router (get_request_language) ↓ AgnoAgentService(lang=lang) ↓ FreeTodoToolkit(lang=lang) ↓ AgnoToolsMessageLoader(lang) ``` ### Adding a New Language 1. Create new directory: `config/prompts/agno_tools/{lang}/` 2. Copy all YAML files from `en/` 3. Translate all messages 4. The loader will automatically detect the new language --- ## 🧪 Testing Tools ### Quick Test Script ```python from lifetrace.llm.agno_tools import FreeTodoToolkit # Test Chinese toolkit_zh = FreeTodoToolkit(lang="zh") print(toolkit_zh.list_todos(status="active", limit=5)) # Test English toolkit_en = FreeTodoToolkit(lang="en") print(toolkit_en.list_todos(status="active", limit=5)) ``` ### Running Tests ```bash uv run python -c " from lifetrace.llm.agno_tools import FreeTodoToolkit tk = FreeTodoToolkit(lang='zh') print(tk.parse_time('明天下午3点')) " ``` --- ## 📊 Observability (Agent Monitoring) The Agno Agent integrates with [Arize Phoenix](https://arize.com/docs/phoenix) + [OpenInference](https://github.com/arize-ai/openinference) for tracing and monitoring. ### Features - **Local JSON Export**: Cursor-friendly trace files for AI analysis - **Phoenix UI**: Optional web-based visualization - **Minimal Terminal Output**: One-line summary per trace ### Configuration In `config/config.yaml`: ```yaml observability: enabled: true # Enable observability mode: both # local | phoenix | both local: traces_dir: traces/ # Trace file directory max_files: 100 # Max files to keep pretty_print: true # Format JSON for readability phoenix: endpoint: http://localhost:6006 project_name: freetodo-agent terminal: summary_only: true # One-line output (recommended) ``` ### Trace File Format Each agent run generates a JSON file in `data/traces/`: ```json { "trace_id": "e078e147372a", "timestamp": "2026-01-23T08:23:48.377470+00:00", "duration_ms": 26910.94, "agent": "breakdown_task", "input": "{\"task_description\": \"Make a video\"}", "output_preview": "Task breakdown:\n1. Define topic...", "tool_calls": [ { "name": "breakdown_task", "args": {"task_description": "Make a video"}, "result_preview": "Task breakdown...", "duration_ms": 26910.94 } ], "llm_calls": [], "status": "success", "span_count": 1 } ``` ### Terminal Output With `summary_only: true`: ``` [Trace] e078e147372a | 1 tools | 26.91s | traces/20260123_082348_e078e147372a.json ``` ### Using Phoenix UI (Optional) ```bash # Start Phoenix server uv run phoenix serve # Access http://localhost:6006 ``` --- ## ✅ Development Checklist When adding new tools: - [ ] Create YAML messages in both `zh/` and `en/` directories - [ ] Create tool mixin class with proper type hints - [ ] Add docstrings for LLM to understand tool usage - [ ] Use `_msg()` for all user-facing messages - [ ] Handle exceptions and return error messages - [ ] Register tool in `tools/__init__.py` - [ ] Add mixin to `FreeTodoToolkit` class - [ ] Register method in `tools` list - [ ] Test with both languages

{ "prompt": "# Agno Agent Development Quick Commands\n\n## Overview\n\nThis guide covers development of **Agno Agent Tools** - the AI-powered todo management toolkit based on the [Agno framework](https://docs.agno.com/).\n\nThe FreeTodoToolkit provides a set of tools for the Agno Agent to manage todos. For the complete list of available tools, please refer to the source code in `llm/agno_tools/tools/` directory.\n\n---\n\n## 🏗️ Architecture\n\n### Directory Structure\n\n```\nlifetrace/\n├── config/prompts/agno_tools/ # Localized messages & prompts\n│ ├── zh/ # Chinese messages\n│ └── en/ # English messages (same structure)\n│\n├── llm/agno_tools/ # Python implementation\n│ ├── __init__.py # Module exports\n│ ├── base.py # Message loader (AgnoToolsMessageLoader)\n│ ├── toolkit.py # Main FreeTodoToolkit class\n│ └── tools/ # Individual tool implementations (organized by category)\n│\n└── observability/ # Agent monitoring (Phoenix + OpenInference)\n ├── __init__.py # Module exports\n ├── config.py # Observability configuration\n ├── setup.py # Initialization entry point\n └── exporters/\n ├── __init__.py\n └── file_exporter.py # Local JSON file exporter\n```\n\n### Design Patterns\n\n- **Mixin Pattern**: Each tool category is a separate mixin class\n- **Composition**: FreeTodoToolkit inherits from all mixins + Agno Toolkit\n- **i18n**: Messages loaded from language-specific YAML files\n- **Lazy Loading**: Database and LLM clients initialized on-demand\n\n---\n\n## 🔧 Adding a New Tool\n\n### Step 1: Add Messages (Both Languages)\n\nCreate or update YAML files in `config/prompts/agno_tools/zh/` and `en/`:\n\n```yaml\n# config/prompts/agno_tools/zh/my_tool.yaml\nmy_tool_success: \"操作成功: {result}\"\nmy_tool_failed: \"操作失败: {error}\"\nmy_tool_prompt: |\n 这是给 LLM 的提示词模板。\n 参数: {param}\n```\n\n```yaml\n# config/prompts/agno_tools/en/my_tool.yaml\nmy_tool_success: \"Operation successful: {result}\"\nmy_tool_failed: \"Operation failed: {error}\"\nmy_tool_prompt: |\n This is a prompt template for LLM.\n Parameter: {param}\n```\n\n### Step 2: Create Tool Mixin\n\nCreate a new file in `llm/agno_tools/tools/`:\n\n```python\n# llm/agno_tools/tools/my_tools.py\n\"\"\"My Tools - Description of what these tools do.\"\"\"\n\nfrom __future__ import annotations\nfrom typing import TYPE_CHECKING\n\nfrom lifetrace.llm.agno_tools.base import get_message\nfrom lifetrace.util.logging_config import get_logger\n\nif TYPE_CHECKING:\n from lifetrace.repositories.sql_todo_repository import SqlTodoRepository\n\nlogger = get_logger()\n\n\nclass MyTools:\n \"\"\"My tools mixin\"\"\"\n\n lang: str\n todo_repo: \"SqlTodoRepository\" # If needed\n\n def _msg(self, key: str, **kwargs) -> str:\n return get_message(self.lang, key, **kwargs)\n\n def my_tool_method(self, param: str) -> str:\n \"\"\"Tool description for LLM to understand when to use it\n\n Args:\n param: Description of the parameter\n\n Returns:\n Result message\n \"\"\"\n try:\n # Implementation\n result = f\"processed {param}\"\n return self._msg(\"my_tool_success\", result=result)\n except Exception as e:\n logger.error(f\"Failed: {e}\")\n return self._msg(\"my_tool_failed\", error=str(e))\n```\n\n### Step 3: Register in Toolkit\n\nUpdate `llm/agno_tools/tools/__init__.py`:\n\n```python\nfrom lifetrace.llm.agno_tools.tools.my_tools import MyTools\n\n__all__ = [..., \"MyTools\"]\n```\n\nUpdate `llm/agno_tools/toolkit.py`:\n\n```python\nfrom lifetrace.llm.agno_tools.tools import (\n ...,\n MyTools,\n)\n\nclass FreeTodoToolkit(\n ...,\n MyTools, # Add mixin\n Toolkit,\n):\n def __init__(self, lang: str = \"en\", **kwargs):\n ...\n tools = [\n ...,\n self.my_tool_method, # Register tool\n ]\n```\n\n---\n\n## 📝 Message Configuration\n\n### YAML Structure\n\nMessages are organized by functionality in `config/prompts/agno_tools/{lang}/` directory. Each YAML file corresponds to a category of messages.\n\n### Message Format\n\n- Use `{placeholder}` for variable substitution\n- Multi-line prompts use YAML `|` syntax\n- Keep messages concise and informative\n\n```yaml\n# Simple message with placeholder\ncreate_success: \"Created todo #{id}: {name}\"\n\n# Multi-line prompt\nbreakdown_prompt: |\n Break down this task into subtasks.\n\n Task: {task_description}\n\n Return JSON format.\n```\n\n### Accessing Messages\n\n```python\n# In tool methods\ndef _msg(self, key: str, **kwargs) -> str:\n return get_message(self.lang, key, **kwargs)\n\n# Usage\nreturn self._msg(\"create_success\", id=123, name=\"Buy groceries\")\n```\n\n---\n\n## 🌐 Internationalization\n\n### Language Selection\n\nLanguage is passed through the call chain:\n\n```\nRequest Header (Accept-Language)\n ↓\nChat Router (get_request_language)\n ↓\nAgnoAgentService(lang=lang)\n ↓\nFreeTodoToolkit(lang=lang)\n ↓\nAgnoToolsMessageLoader(lang)\n```\n\n### Adding a New Language\n\n1. Create new directory: `config/prompts/agno_tools/{lang}/`\n2. Copy all YAML files from `en/`\n3. Translate all messages\n4. The loader will automatically detect the new language\n\n---\n\n## 🧪 Testing Tools\n\n### Quick Test Script\n\n```python\nfrom lifetrace.llm.agno_tools import FreeTodoToolkit\n\n# Test Chinese\ntoolkit_zh = FreeTodoToolkit(lang=\"zh\")\nprint(toolkit_zh.list_todos(status=\"active\", limit=5))\n\n# Test English\ntoolkit_en = FreeTodoToolkit(lang=\"en\")\nprint(toolkit_en.list_todos(status=\"active\", limit=5))\n```\n\n### Running Tests\n\n```bash\nuv run python -c \"\nfrom lifetrace.llm.agno_tools import FreeTodoToolkit\ntk = FreeTodoToolkit(lang='zh')\nprint(tk.parse_time('明天下午3点'))\n\"\n```\n\n---\n\n## 📊 Observability (Agent Monitoring)\n\nThe Agno Agent integrates with [Arize Phoenix](https://arize.com/docs/phoenix) + [OpenInference](https://github.com/arize-ai/openinference) for tracing and monitoring.\n\n### Features\n\n- **Local JSON Export**: Cursor-friendly trace files for AI analysis\n- **Phoenix UI**: Optional web-based visualization\n- **Minimal Terminal Output**: One-line summary per trace\n\n### Configuration\n\nIn `config/config.yaml`:\n\n```yaml\nobservability:\n enabled: true # Enable observability\n mode: both # local | phoenix | both\n local:\n traces_dir: traces/ # Trace file directory\n max_files: 100 # Max files to keep\n pretty_print: true # Format JSON for readability\n phoenix:\n endpoint: http://localhost:6006\n project_name: freetodo-agent\n terminal:\n summary_only: true # One-line output (recommended)\n```\n\n### Trace File Format\n\nEach agent run generates a JSON file in `data/traces/`:\n\n```json\n{\n \"trace_id\": \"e078e147372a\",\n \"timestamp\": \"2026-01-23T08:23:48.377470+00:00\",\n \"duration_ms\": 26910.94,\n \"agent\": \"breakdown_task\",\n \"input\": \"{\\\"task_description\\\": \\\"Make a video\\\"}\",\n \"output_preview\": \"Task breakdown:\\n1. Define topic...\",\n \"tool_calls\": [\n {\n \"name\": \"breakdown_task\",\n \"args\": {\"task_description\": \"Make a video\"},\n \"result_preview\": \"Task breakdown...\",\n \"duration_ms\": 26910.94\n }\n ],\n \"llm_calls\": [],\n \"status\": \"success\",\n \"span_count\": 1\n}\n```\n\n### Terminal Output\n\nWith `summary_only: true`:\n\n```\n[Trace] e078e147372a | 1 tools | 26.91s | traces/20260123_082348_e078e147372a.json\n```\n\n### Using Phoenix UI (Optional)\n\n```bash\n# Start Phoenix server\nuv run phoenix serve\n\n# Access http://localhost:6006\n```\n\n---\n\n## ✅ Development Checklist\n\nWhen adding new tools:\n\n- [ ] Create YAML messages in both `zh/` and `en/` directories\n- [ ] Create tool mixin class with proper type hints\n- [ ] Add docstrings for LLM to understand tool usage\n- [ ] Use `_msg()` for all user-facing messages\n- [ ] Handle exceptions and return error messages\n- [ ] Register tool in `tools/__init__.py`\n- [ ] Add mixin to `FreeTodoToolkit` class\n- [ ] Register method in `tools` list\n- [ ] Test with both languages", "variables": [], "notes": "This guide covers development of **Agno Agent Tools** - the AI-powered todo management toolkit based on the [Agno framework](https://docs.agno.com/).", "metadata": { "source": { "provider": "github-cursor-commands", "repository": "https://github.com/zzw4257/free-to-do-app", "path": ".cursor/commands/agno_agent.md", "ref": "4312e3ff40b6f4f396d26e9b4cc41a38e41fae3f", "url": "https://github.com/zzw4257/free-to-do-app/blob/4312e3ff40b6f4f396d26e9b4cc41a38e41fae3f/.cursor/commands/agno_agent.md", "key": "zzw4257/free-to-do-app/.cursor/commands/agno_agent.md" } } }

agno_agent

About