Answers only with typos and typed with your little finger, I 'll let you guys start:
Python, SQL, Docker, Terraform, GCP Cloud and Snowflake. Related to AI: Vertex AI, Agent2Agent protocols, Claude Code and Cortex
I started in Notepad++, then moved to pycharm. When I started coding for pyRevit I was using ChatGPT to write the first iteration, then copy+paste back and forth (this was befor any AI in IDE was a useful thing) or I did minor manual fixes.
Started using Cursor about 2-2.5years ago, and still use it. I now mainly rely on rules (both user/team and project related), docs and some skills as a harness to eliminate typical errors and improve outcome. I also use MCP (usually custom made or very project related) to some extent, but very limited.
For larger projects I first plan with ChatGPT or Claude. When Im happy with the plan I ask for a json-promt that i feed to Cursor to build the project. I give it the json-promt as a file, ask it to make some .md to reference as it builds (plans, documentation of the project and a userguide if needed). This often gives me a usable MVP in a very short time, when starting new projects.
I want to look into custom memory to see if it has benefits, just havent gotten around to testing it.