repo for AI hackathon
For this version, I moved beyond just a single test tool and built out a proper CLI workflow so I can interact with GitHub more naturally.
Here’s what changed:
-
Added a bunch of new tools:
search_repositories→ lets me search repos by keywordget_me→ grabs my own GitHub profile infoget_team_members→ pulls members from an org teamlist_issues→ shows all issues in a repolist_prs→ shows pull requestslist_discussions→ shows active repo discussions
-
Built a small context repo feature:
set context repo <repo>→ so I don’t have to keep typing the repo name again and againreset context repo→ clears it if I want to switch
-
Tested it end-to-end in the CLI:
- user info fetch works
- repo search returns correct results
- issues, PRs, and discussions all list out cleanly
- context repo made commands way smoother
📸 Screenshots I’ll add here later:
This version feels like the first time the project is actually usable — I can run real GitHub queries from the CLI instead of just testing one endpoint.
-
Install the dependencies:
pip install -r requirements.txt
-
To start the CLI, just run:
python main.py
- ✅ Built initial Haystack MCP agent
- 🛠️ Tool:
get me - 🧪 Tested basic NLP query: "Can you retreive my user data from the github"
- 📸 Screenshot:

- Steps to run: just run the "starter_code.py file"
This version lays the foundation for future GitHub-aware NLP capabilities.






