A seven-agent study workflow for turning a GATE Computer Science topic into a structured learning package with theory, previous-year questions, worked examples, pattern analysis, practice questions, and mathematical verification.
Stage: Recovered architecture blueprint only. No executable n8n workflow is included because the original export was not available on this laptop.
- Course Planner
- Theory Agent
- Previous-Year Question Retrieval Agent
- Solved-Example Agent
- Problem-Pattern Agent
- Practice Generator
- Mathematics Verifier
The orchestrator combines all outputs into a single final_markdown study unit while preserving citations, uncertainty, and correction notes.
The original design targeted n8n and the OpenAI Responses API. Each agent receives a scoped instruction and a shared course context; the mathematics verifier runs after content generation and can return corrections before final assembly.
Reconstructed from surviving ChatGPT project memory. The original n8n export was not available on this laptop, so this repository documents the workflow honestly instead of presenting invented JSON as the recovered automation.
See docs/workflow.md for interfaces and evaluation gates.
Baavish L. — Computer Science Engineering student and AI systems builder.
Copyright © 2026 Baavish L. All rights reserved.