DietrichGebert/ponytail ? reverse-engineered prompt
Reverse engineered prompt
Build me a project called Ponytail, a plugin or ruleset that makes AI coding agents behave like the laziest senior dev in the room. I want it to guide the agent to first ask whether the code even needs to exist, then prefer the standard library, built in browser or platform features, existing dependencies, or a one line solution before writing custom code.
It should work with the main coding agents this repo supports, with clear install instructions and the right adapters or rule files for each. Where the host supports commands, add simple mode switching like lite, full, ultra, and off, plus commands for review, audit, help, and debt tracking. If Ponytail takes a shortcut, it should leave a clear ponytail: comment in the code so there is an obvious upgrade path later.
Please include examples, docs, tests, and benchmark scripts that show the whole point, less code, faster output, and lower cost. Keep it opinionated but not reckless, so security, accessibility, validation, and data loss handling are never skipped. Look up current agent docs online if you need to.
Are you gonna build this?
make sure you review the code using coderabbit