Five Claude Code skills that feel illegal to know. They are not. Each one is a real operating mode wearing a name you will actually remember: a Mandarin compressor for agent fleets, an anti-slop design enforcer, a six-agent tournament with a judge, a fact-checker that attacks its own answers, and a startup analyst with Euro Summer energy. The name is the bit; the output is the point.
These are standard Claude Code skills. Drop any folder into your skills directory and invoke it with its slash command.
# 1. Clone the repo
git clone https://github.com/Jakeschincariol/claude-illegal-codes.git
cd claude-illegal-codes
# 2. Copy any skill folder into your Claude Code skills directory
cp -r crash-out ~/.claude/skills/
# 3. Use it inside Claude Code by typing its slash command
# /crash-outOr take all five at once:
cp -r china-maxing aura-farming 67 crash-out larp ~/.claude/skills/Each skill lives in its own folder containing a single SKILL.md. No time to install? Paste the body of any SKILL.md at the top of a chat as a mode and you get most of the effect.
| Command | What it does |
|---|---|
/china-maxing |
Translates every sub-agent prompt into Mandarin before dispatch. Fewer tokens, same instructions, smaller bill. |
/aura-farming |
Maximum-aesthetic mode for ads, decks, and pages. Bans every AI-slop tell and audits the design before it ships. |
/67 |
Six agents solve your task six different ways. A seventh judges, ranks, and steals the strongest parts of each. |
/crash-out |
Claude becomes obsessed with proving itself wrong. Every claim gets attacked; only survivors ship. |
/larp |
Analyzes any startup idea as a 24-year-old founder who raised $100M for a ChatGPT wrapper, then builds the strongest real business model underneath the bit. |
/china-maxingsavings depend on the tokenizer and the prompt. Prose-heavy briefs compress hardest; code does not compress at all, so the skill leaves code alone./67costs six times the agent runs. Use it where quality pays for itself.- "Illegal" is a joke. Everything here is prompt engineering.
Made by Jake Schincariol, opusjake.ai. Read the full write-up.
MIT