Skip to content

STUDENT_ONBOARDING.md

Latest commit

 

History

History
111 lines (78 loc) · 2.93 KB

STUDENT_ONBOARDING.md

File metadata and controls

111 lines (78 loc) · 2.93 KB
Raw

Student Onboarding — AgentStore

Welcome to AgentStore! This guide helps you get set up and start contributing.

1. Clone the Repository

git clone https://github.com/TechX-Resources/AgentStore.git
cd AgentStore

2. Choose Your Track

Track You'll Work In Example First Tickets
LLM agents/, tools/, backend runner Agent manifests, tool traces, prompt templates
Data Science data-science/, datasets/ Trending algorithm, rating analysis, mock data
Full Stack frontend/, backend/ Agent cards UI, API endpoints, detail pages

3. Set Up Your Environment

Backend

cd backend
python -m venv venv
# macOS/Linux:
source venv/bin/activate
# Windows:
venv\Scripts\activate

pip install -r requirements.txt
uvicorn app.main:app --reload

Visit http://127.0.0.1:8000/docs to see the API placeholder.

Frontend

cd frontend
npm install
npm run dev

Visit http://localhost:5173 to see the UI placeholder.

Data Science

cd data-science
pip install -r requirements.txt

Open notebooks in data-science/notebooks/ or run scripts in data-science/analytics/.

4. Explore the Codebase

Start with these files:

  1. PROJECT_OVERVIEW.md — What we're building
  2. docs/architecture.md — How the system is organized
  3. docs/user-stories.md — What users need
  4. agents/manifests/ — Sample agent definitions
  5. tools/manifests/ — Sample tool definitions
  6. kanban/ — Available tickets

5. Pick a Ticket

  1. Go to the Kanban board (link in kanban/README.md).
  2. Pick a ticket labeled for your cohort.
  3. Move it to "In Progress" and assign yourself.

6. Create a Branch

git checkout main
git pull origin main
git checkout -b feature/AS-XX-short-description

7. Implement and Submit

  1. Make your changes within the ticket scope.
  2. Test locally (backend, frontend, or scripts as relevant).
  3. Commit with a clear message.
  4. Push and open a pull request.
  5. Reference the ticket ID in your PR.

Rules of the Road

  • Skeleton first — Many files have TODO comments; implement only what your ticket asks for.
  • No secrets — Never commit API keys or .env files.
  • No scope creep — One ticket, one PR.
  • Ask questions — Use GitHub Discussions or ask your cohort lead.

Key Concepts

Agent Manifest

A JSON file describing an agent: name, tools, permissions, inputs/outputs.

Tool Manifest

A JSON file describing a tool: name, schema, permissions, mock output notes.

Tool-Call Trace

A step-by-step log of what happens when an agent runs (identify tools → call tool → process response → generate answer).

Mock Storage

The backend starts with JSON file storage — no database required for MVP.

Good luck — build something awesome!