AI tools: professional (Cursor & Gemini)
This course covers a professional workflow with two agentic tools: Cursor, an IDE, and Gemini CLI, an open-source terminal agent. It explains how Cursor decides which rules apply to a file, how nested AGENTS.md files combine, and why a rules file named MEMORY.md does not load automatically. It covers Gemini CLI's built-in tools, GEMINI.md and non-interactive mode. It also covers Plan Mode, checking a context file against the code it describes, and reviewing work from cloud and parallel agents.
What you'll learn
- ✓How Cursor's alwaysApply, globs and description fields decide when a rule applies
- ✓Why a rules file's filename decides whether a tool loads it automatically
- ✓Using Gemini CLI in interactive and non-interactive mode
- ✓Reviewing and editing a plan in Cursor's Plan Mode before any file changes
- ✓Checking a context file's claims against the code it describes
- ✓Reviewing branches from Cursor's Cloud Agents and parallel worktree agents
From one loop to a professional, multi-tool workflow
"AI tools: introduction (Cursor & Antigravity)" named the loop underneath every agentic tool: read a file, write a change, run a command, read the result, repeat. Its example had an agent add published? to an Article model: open app/models/article.rb, write the method and run the test suite. "AI…
Read the full lesson free →Curriculum
7 lessonsFrom one loop to a professional, multi-tool workflow
Introduces Cursor and Gemini CLI as the same agent loop in two working contexts, and what professional means here.
Cursor's rules system, past the basics
Covers Cursor's four-branch rule logic, User Rules limits, nested AGENTS.md precedence, and why MEMORY.md does not auto-load.
Gemini CLI: an agent that lives in your terminal
Presents Gemini CLI's built-in tools, its GEMINI.md context file, and interactive and non-interactive modes for scripts and CI.
Planning before touching a single file
Explains Cursor's Plan Mode: clarifying questions, codebase research and an editable plan reviewed before any file changes.
Giving either tool this project's context
Checks MEMORY.md's stale Icons section against LucideHelper::ICONS, the star fallback, and the commit that fixed the file.
Cloud and parallel agents: reviewing, not typing along
Covers Cursor's Cloud Agents, parallel agents in Git worktrees, Gemini CLI checkpointing, and reviewing finished branches like pull requests.
Putting it together: context, a plan, and a review, every time
Ties together checking rules files, reviewing plans early, and reviewing unattended work with the rigor of a pull request.