happyrust

happyrust / Warp orchestrator multi-agents rule

Created October 28, 2025 11:20 — forked from tao12345666333/Warp orchestrator multi-agents rule

	You have three sub-agents available: amp, auggie, and gemini. You can break down the tasks that need to be executed and assign them separately to these agents.
	Specifically, use `amp -x "your task description"`, `auggie -p "task description"`, and `gemini -p "task description"`. Once they complete their respective tasks, you can collect the results and perform the final review and summary.

happyrust / AI_007_prompt.en.txt

Created October 21, 2025 14:12 — forked from lexrus/AI_007_prompt.en.txt

AI 通宵干活 prompt

	__This is the requirements section, please replace it yourself__

	Above are the requirements, below are the decision-making requirements:
	- I'm going to sleep now, all product and technical decisions are under your control, I'll come to review the results tomorrow morning
	- I'll sleep for about 8 hours, so don't rush to produce results, take your time with all the details, don't save tokens, getting it right is most important
	- When you encounter uncertainties, first check the documentation in context7, if still uncertain, search online
	- Execute all subtasks in phases, run unit tests after each phase is completed, to avoid having to start over after completing everything only to find it doesn't work
	- Don't get hung up on the UI part, try to use mainstream conventional approaches, I'll polish the details tomorrow, you just focus on getting the functionality working