open-ptc-agent

🥈Silver

Open PTC Agent automates code execution with Programatic Tool Calling (PTC) for operations teams. It connects to LangChain and Langraph workflows, enabling developers to build custom agents for Claude LLM. The open-source tool is written in Python 3.12+ and is licensed under MIT.

700730Updated Today

Intermediate30min to implementautomation

Saves ~240 min per use

Quick InstallView Source

git clone https://github.com/Chen-zexi/open-ptc-agent.git

Works with:

Claude

Overview

About This Skill

https://github.com/Chen-zexi/open-ptc-agent

How to Use

["Install the Open PTC Agent via pip: `pip install open-ptc-agent` (requires Python 3.12+). Verify the installation with `ptc-agent --version`.","Configure the agent by creating a Python script or Jupyter notebook. Set up the workflow (LangChain/LangGraph) and define the model parameters (e.g., `model='llama3'`, `temperature=0.7`).","Define the task using the `execute()` method. Specify the task description, required tools (e.g., `pandas`, `requests`), and input data (e.g., file paths, API endpoints).","Run the agent via command line (`ptc-agent run --script your_script.py`) or directly in your IDE. Monitor logs for progress and errors.","For debugging, enable verbose logging (`--verbose` flag) or use the agent's built-in error handling to retry failed tasks. Review the output directory for generated artifacts."]

Use Cases

Automate the analysis of financial market data by fetching and processing stock prices for multiple tickers.

Create complex data visualizations by executing Python code that aggregates and transforms raw data.

Streamline the generation of reports by automating the extraction and formatting of data from various sources.

Implement background tasks that allow for asynchronous execution of multiple data processing jobs without blocking the main workflow.

Setup & Installation

Quick Install

No install command available. Check the GitHub repository for manual installation instructions.

Alternative Install (Git Clone)

git clone https://github.com/Chen-zexi/open-ptc-agent

Requirements

Claude Code or compatible AI agent
Works with: Claude

Quick Start Guide

Install the Skill

Copy the install command above and run it in your terminal.

Open Your AI Agent

Launch Claude Code, Cursor, or your preferred AI coding agent.

Try It Out

Use the prompt template or examples below to test the skill.

Customize

Adapt the skill to your specific use case and workflow.

Usage Examples

Prompt Template

Use the Open PTC Agent to [TASK] by integrating it with [LANGCHAIN/LANGGRAPH_WORKFLOW]. Configure the agent with [PYTHON_CONFIGURATION] and execute it via [COMMAND_LINE_INTERFACE]. Monitor the output and handle any errors using [LOGGING/DEBUGGING_TOOLS].

Example Output

```python
# Example: Automating a data processing task with Open PTC Agent
from open_ptc_agent import PTC_Agent
from langchain_community.llms import Ollama

# Initialize the agent with a custom workflow
agent = PTC_Agent(
    workflow="data_processing_pipeline",
    config={
        "model": "llama3",
        "temperature": 0.7,
        "max_tokens": 1000
    }
)

# Define the task: Process a CSV file and generate a summary report
result = agent.execute(
    task="Process the sales_data.csv file to generate a summary report with total revenue, top products, and sales trends.",
    tools=["pandas", "matplotlib"],
    input_data={
        "file_path": "/data/sales_data.csv",
        "output_dir": "/reports"
    }
)

# Output the results
print("Task completed successfully!")
print("Generated report:", result['report_path'])
print("Summary metrics:", result['metrics'])
```

**Execution Log:**
```
[2024-05-20 14:30:22] INFO: Initializing PTC Agent with workflow 'data_processing_pipeline'
[2024-05-20 14:30:25] INFO: Loading model 'llama3' with config {'temperature': 0.7, 'max_tokens': 1000}
[2024-05-20 14:30:28] INFO: Executing task: Process sales_data.csv
[2024-05-20 14:31:15] INFO: Data loaded successfully. Rows: 12,450
[2024-05-20 14:32:03] INFO: Generating summary report...
[2024-05-20 14:32:45] INFO: Report saved to /reports/sales_summary_2024-05-20.pdf
[2024-05-20 14:32:45] INFO: Metrics calculated: Total Revenue: $1,245,678.90, Top Product: Widget Pro (34% of sales)
```

The agent successfully processed the CSV file, generated a visual summary report with matplotlib, and calculated key metrics. The report is now ready for review in the specified output directory.