Find agent skills by outcome

The game should feel lively

slightly sarcastic

- Local/open-source (via Ollama/LM Studio/etc.): MythoMax

DeepSeek V3

- Gemini (Google) — Gemini 1.5 / 2.0 family: Fast

multimodal potential

- Claude (Anthropic) — Claude 3.5 Sonnet / Claude 4: Top-tier rule adherence

nuanced humor

This prompt performs best on models with strong long-context handling (≥128k tokens preferred)

precise instruction-following

"YOU PROBABLY DON'T KNOW THIS" Game

instruction: Reinforce fixed camera: no zoom

no pan

instruction: Include photorealistic

instruction: Mention low/high-key lighting

warm/cold sources

- Think in diffs

not files.

Deviation from this style is a correctness issue

not a preference issue.

war

Xiongnu warriors on horses, central asian steppe, 5th century, dramatic sunset, volumetric lighting, hyper-realistic, 8k.

- If no structure exists

propose one and wait for approval.

- Use modular

responsibility-focused files.

8. **Prefer minimal

incremental changes.**

If the answer is anything other than append-only

the agent must STOP and ask for confirmation.

- Am I rewriting for clarity

cleanup

Before modifying any md file

the agent must internally verify:

- Until approval is given

limit output to analysis

- Phrases like “go ahead”

“implement”

- **All state

memory

(e.g. Antigravity install folder

home directories

- The agent must not make product

- When instructions conflict

the following precedence applies: