Three-year RPG dev shares workflow: make Claude test what actually changed
tschilpi · reddit · 2026-11-02
A developer who spent three years building Lore & Legends Maker, a fantasy RPG with an AI Dungeon Master, shares the most valuable lesson from developing it with Claude Code and Codex: give the AI more ways to inspect results instead of trusting its reports.
- Visual checks: dedicated inspection pages for spells, cards, and maps let Claude capture frames with Playwright, inspect them, and iterate before human review.
- Gameplay bugs: Claude gets a recorded player session and compares actions, inventory, and state changes against the narration — e.g., a merchant accusing the player of drawing steel when they only own a makeshift stick.
- Core challenge: the game maintains world state while the AI interprets actions and narrates; keeping those aligned is most of the work, and a successful build doesn't reveal whether an NPC reacted to something that never happened.
The game is free to try at llmsaga.com.
More from coding & agent
- Perplexity open-sources WANDR: 500 research tasks, 170,495 evidence points, URL-rechecking grader — thisguyknowsai · 2026-11-02
- 4 AIs, 96 questions, 2,671 citations: 74% are blog posts, 99% on companies' own sites — tibo_maker · 2026-11-02
- Agent audit said 0 blocks while hooks logged 7: Claude Code transcript drift broke a blind detector — Ok-Motor-9812 · 2026-11-02
- Tencent unveils Hyra 2.0, a research agent that builds reusable scientific tools — TencentHunyuan · 2026-11-02
- What broke when I ran a leaderless agent swarm with no orchestrator — 7 failure modes and fixes — LostHumanity67 · 2026-11-02
- Forced yes/no prompts push GPT and Claude to 95%+ confidence on Goldbach's unproved halt test — Nir777 · 2026-11-02