---
title: "Rule-citing AI dungeon master game"
canonical: "https://agenticfeats.com/c/rule-citing-ai-dungeon-master-game-custom-agent/"
agents: ["Custom agent"]
evidence_grade: "A"
outcome: "completed"
job_type: "Coding"
roles: ["Developer"]
source_url: "https://dev.to/prime_299792/the-goblin-warren-an-ai-dungeon-master-that-never-invents-a-rule-4d66"
evidence_url: "https://d-and-d-one.vercel.app/"
source_date: "2026-10-04"
published: "2026-10-13"
updated: "2026-10-13"
---

> From Agentic Feats: sourced examples of AI agents at work. Map of the site for agents: https://agenticfeats.com/llms.txt

# Rule-citing AI dungeon master game

The author built a browser D&D 5e dungeon crawl whose AI Dungeon Master answers rules questions by looking up and citing rules stored in Sanity.

Result: A live game with 87 rules, 33 conditions, 14 spells and 15 monsters in two editions, answering rules questions in about 1.3 to 2.2 seconds.

## How it was set up

Sanity dataset seeded with SRD data, two Context MCP endpoints (knowledge base and GROQ), DeepSeek on Baseten via AI SDK. Server prefetches lookups before the model runs.

## Worth knowing

The MCP endpoints need an org token with the Context Viewer role. A project token returns a 403 error, which cost the author an hour.

## Evidence note

Live demo, GitHub repo, screenshots and GIFs

## Sources

- Original: https://dev.to/prime_299792/the-goblin-warren-an-ai-dungeon-master-that-never-invents-a-rule-4d66
- Evidence: https://d-and-d-one.vercel.app/

## The receipt

- Agent: Custom agent
- Evidence: Shown (grade A, artifact linked)
- Outcome: Completed
- Job type: Coding
- Roles: Developer
- Source: Blog
- Published: 2026-10-04
- Added: 2026-10-13