feat: RAG knowledge base has no pre-loaded regulatory documents — users must add their own PDFs to get answers about EU AI Act, GDPR, and ISO 42001, but the README implies these are built-in

[prince-pokharna]

🐛 Problem Statement

The AegisAI roadmap lists: "Pre-loaded regulatory knowledge base (EU AI Act PDF, GDPR, ISO 42001, NIST AI RMF)" as a planned feature. The README's RAG module description states: "Ask natural language questions about EU AI Act, GDPR, ISO 42001 — grounded answers from regulatory source docs."

However, backend/data/regulatory_docs/ (where RAG source documents should live) currently contains only a placeholder instruction: "Add your regulatory PDFs here." This means a fresh AegisAI deployment cannot answer any regulatory question out of the box — despite the README explicitly advertising EU AI Act Q&A as a core feature.

A user who installs AegisAI following the Quick Start guide and asks "What are the requirements for high-risk AI systems under the EU AI Act?" will get "I don't have relevant information in my knowledge base" — directly contradicting the README.

Proposed Fix

Add a scripts/download_regulatory_docs.py script that downloads publicly available regulatory documents from official sources and places them in backend/data/regulatory_docs/:

# scripts/download_regulatory_docs.py
"""
Downloads publicly available regulatory documents for AegisAI's RAG knowledge base.
All documents are freely available from official government/standards sources.
"""

import httpx
from pathlib import Path

REGULATORY_DOCS = [
    {
        "name": "EU_AI_Act_2024.pdf",
        "url": "https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=OJ:L_202401689",
        "description": "EU AI Act — Official Journal of the European Union (2024/1689)",
    },
    {
        "name": "NIST_AI_RMF_1_0.pdf",
        "url": "https://nvlpubs.nist.gov/nistpubs/ai/NIST.AI.100-1.pdf",
        "description": "NIST AI Risk Management Framework 1.0",
    },
    # GDPR — available from EUR-Lex
    # ISO 42001 — note: ISO docs are not freely available; use the publicly released synopsis
]

def download_regulatory_docs():
    output_dir = Path("backend/data/regulatory_docs")
    output_dir.mkdir(parents=True, exist_ok=True)

    for doc in REGULATORY_DOCS:
        output_path = output_dir / doc["name"]
        if output_path.exists():
            print(f"Already downloaded: {doc['name']}")
            continue
        print(f"Downloading {doc['name']}...")
        response = httpx.get(doc["url"], follow_redirects=True, timeout=60)
        output_path.write_bytes(response.content)
        print(f"Saved to {output_path}")

if __name__ == "__main__":
    download_regulatory_docs()

Add this to docker-compose.yml as a one-time initialization step, and document it in docs/getting-started.md as a required setup step.

Files to Create/Modify

File	Change
scripts/download_regulatory_docs.py	New download script
docker-compose.yml	Add init container that runs the script on first startup
docs/getting-started.md	Document the script as a required setup step
backend/data/regulatory_docs/README.md	Update placeholder with actual instructions

Suggested labels: enhancement, rag, roadmap, level: beginner

I would like to work on this. Could you please assign it to me?

[prince-pokharna]

Hey @SdSarthak ,
Please assign it to me under GSSOC 2026.
/assign /claim