A Haystack‑powered MCP client that automates common GitHub workflows: typo detection, issue searching, and forking repositories.
Submitted to the Spring Into Haystack challenge, this agent demonstrates four static example queries to validate core functionality within scope.
NOTE: Read the original challenge description here: https://github.com/deepset-ai/spring-into-haystack
The agent is designed to be user-friendly, with a simple command-line interface. Users can select from pre-existing queries, and the agent will respond with the appropriate actions or information.
The intent of the project is to only demonstrate the core functionality of the agent. The static examples are designed to validate the agent's capabilities without risking real-world issues or changes to repositories that may arise from LLM hallucinations. This approach ensures that the agent's core functionalities are showcased effectively while maintaining a safe and controlled environment.
Each example shows the user input, the agent's output, and the corresponding Github changes (if any).
Input
Output
GitHub Changes
Input
Output
Input
Output
Input
Output
GitHub Changes
-
Clone the repository
git clone https://github.com/YourUser/themimikyu-spring-into-haystack.git cd themimikyu-spring-into-haystack
-
Install dependencies
pip install -r requirements.txt
-
Environment Variables
- Copy
.env.template
to.env
- Fill in your
GITHUB_PERSONAL_ACCESS_TOKEN
andOPENAI_API_KEY
- Copy
-
Setup GitHub MCP Server (requires Docker)
docker --version # verify Docker is installed docker pull ghcr.io/github/github-mcp-server
-
Run the Agent
python src/github-agent.py
This project is a derivative of the Spring Into Haystack event and licensed under the same license (MIT License). See the LICENSE file for details.