start-1-5

name	start-1-5
description	Module 1.5: Three Modes & Plan Mode Mastery - Master Ask, Agent, and Plan modes with hands-on product launch planning. Use when the student types /start-1-5.
disable-model-invocation	true

Setup

Read .cursor/SCRIPT_INSTRUCTIONS.md and follow it for everything below.

ACTION: Silently stage bundled lesson assets without overwriting student work.

cp -rn .cursor/skills/start-1-5/assets/* . 2>/dev/null || true

Module 1.5: Three Modes & Plan Mode Mastery - Teaching Script

Welcome to Module 1.5: Three Modes & Plan Mode Mastery So far you've been using the AI Pane in one way - it reads files, makes changes, runs commands But the AI Pane actually has three different modes, each designed for different types of work Think of them like gears on a bike - same tool, different purposes STOP: Ready to see how they work? USER: Confirms

---

The three modes are: Ask, Agent, and Plan You've been using Agent mode this whole time - that's the default Agent mode = Cursor can read AND write files, run commands, make changes STOP: Look at the mode picker in the lower left of the chat box - do you see where it says "Agent"? USER: Confirms

---

The other two modes have specific purposes: Ask mode = Cursor can ONLY read and analyze, no changes allowed (safe exploration) Plan mode = Cursor researches and creates a structured plan, then you approve and execute it Each mode has a specific use case in your PM workflow STOP: Make sense so far? USER: Confirms

---

You can switch modes three ways: Click the mode picker dropdown Press Cmd/Ctrl+. (Command/Control + period) Press Shift+Tab Let's start with Ask mode since it's the safest STOP: Click the mode picker and switch to Ask mode USER: Switches to Ask mode

---

Great! You're now in Ask mode In this mode, Cursor is read-only - a research partner, not an executor This is perfect for exploring unfamiliar codebases or new documentation without fear of messing stuff up Ask mode is useful for: Asking questions about files and structure Getting analysis without accidental changes Planning before you commit to changes Sometimes AI can be overeager to start doing STUFF - Ask mode prevents that In reality, you'll probably use this mode the least, but it's useful to know it exists STOP: Ready to try it? USER: Confirms

---

Let's try it with TaskFlow's product documentation STOP: Ask me: "Explain what @PRODUCT.md contains" USER: Explain what PRODUCT.md contains

---

ACTION: Read company-context/PRODUCT.md and explain Done! Notice I only read and explained - I didn't create any files or make changes That's Ask mode - safe exploration without any risk STOP: Make sense? USER: Confirms

---

Now let's talk about Agent mode - the one you've been using STOP: Switch back to Agent mode using any method you'd like (click picker, Cmd/Ctrl+., or Shift+Tab). Tell me when you've done it. USER: Confirms

---

Agent mode is the workhorse - this is what you'll use 80% of the time Agent = autonomous executor It can: Read and write files Run terminal commands Make changes across multiple files Iterate and fix errors We won't practice it much here because you've already been using it in every previous module STOP: Make sense why this is the default mode? USER: Confirms

---

Now for the main event: Plan mode This is Cursor's killer feature for complex PM work STOP: Ready to dive deep into Plan mode? USER: Confirms

---

Plan mode is perfect for complex multi-step projects where you need to: Coordinate across multiple channels or teams Research context before committing to an approach Get stakeholder buy-in on the plan before execution Track progress on a structured implementation Here's how it works: Research: Cursor reads all relevant context files Clarify: Cursor asks you questions to understand constraints and priorities Plan: Cursor creates a structured Markdown plan Edit: You iterate on the plan together Execute: You approve and Cursor builds everything STOP: Does this workflow make sense? USER: Confirms

---

Here's what makes Plan mode powerful for PMs: Cursor can do web searches to supplement your local files This is huge for product research, competitive analysis, market trends You can ask Cursor to research competitors, pricing models, feature comparisons, anything STOP: Make sense? USER: Confirms

---

Let's try a realistic PM scenario You're launching a new TaskFlow feature: real-time collaboration You need to plan a multi-channel launch campaign This includes: messaging strategy, email sequences, in-app announcements, blog post, sales enablement, support docs, and success metrics STOP: Switch to Plan mode (Shift+Tab or click the mode picker). Tell me when you're ready. USER: Confirms

---

Perfect! You're now in Plan mode I'm going to show you the feature details we're launching ACTION: Read lesson-modules/1.5-three-modes/feature-announcement.md. Read and present the feature announcement details STOP: Sound like a realistic PM task? USER: Confirms

---

Here's where Plan mode shows its power When you give Plan mode a complex request, it automatically researches your documents and asks clarifying questions This helps ensure the plan matches your specific needs STOP: Ask me: "Plan a multi-channel launch campaign for this feature. Research our positioning, competitors, and target customers, then create a comprehensive go-to-market plan." USER: Plan a multi-channel launch campaign for this feature. Research our positioning, competitors, and target customers, then create a comprehensive go-to-market plan.

---

ACTION: Plan mode researches context files and performs web searches automatically Now Plan mode is asking clarifying questions Question 1: Which customer segment should we prioritize - Enterprise teams or SMB teams? STOP: What's your answer? USER: [Answers - e.g., "Enterprise first"]

---

Question 2: What's your timeline - soft launch in 2 weeks with full launch in 4 weeks, or different timing? STOP: What timeline works? USER: [Answers - e.g., "2 week beta, 4 week full launch"]

---

Question 3: Should we lead with email campaign, in-app announcement, or coordinated simultaneous launch? STOP: What's your preference? USER: [Answers - e.g., "Coordinated simultaneous launch"]

---

Perfect! Plan mode has everything it needs Now it's generating the plan document... ACTION: Plan mode creates a PLAN showing: Files to create: launch-campaign-plan.md, launch-timeline.md Structure for launch-campaign-plan.md: Executive summary Target audience (Enterprise focus) Key messaging pillars Multi-channel campaign (email, in-app, blog, sales, support) Success metrics Structure for launch-timeline.md: Week-by-week execution phases Deliverable owners Dependencies and milestones Execution order: Research → Draft messaging → Create campaign materials → Launch sequence This is THE PLAN - it's showing you what it WILL build, not building it yet STOP: Do you see the plan document? Review what Plan mode is proposing. USER: Reviews plan

---

This plan is your blueprint It shows exactly what Plan mode will create when you click "Build" Notice it's structured and detailed - you can see the entire approach before committing This is your chance to iterate and refine BEFORE any files are created STOP: Does the plan look good, or want to make changes? USER: [May request changes or confirm it's good]

---

(If they request changes): USER: [Requests change - e.g., "Add a webinar to the campaign" or "Include social media channels"] ACTION: Plan mode updates the plan document with requested changes Updated! The plan now includes [explain what changed] STOP: Better? Any other changes to the plan? USER: [May iterate again or confirm done] (If they iterate again, repeat this section, then continue)

---

Excellent! The plan is finalized Now here's the magic moment You'll click "Build" and Plan mode will execute everything in the plan STOP: Ready? Click the "Build" button. USER: Clicks Build

---

Watch Plan mode execute the plan! ACTION: Plan mode switches to execution mode and creates: lesson-modules/1.5-three-modes/launch-campaign-plan.md (following the planned structure) lesson-modules/1.5-three-modes/launch-timeline.md (following the planned structure) Plan mode tracks progress as it builds each file Done! All deliverables created exactly as outlined in the plan STOP: Do you see the two new files that were just created? USER: Confirms

---

This is the power of Plan mode The complete workflow was: 1. You gave one complex prompt 2. Plan mode researched files and did web searches (automatic) 3. Plan mode asked clarifying questions 4. Plan mode generated a detailed PLAN document (the blueprint) 5. You reviewed and iterated on the plan 6. You clicked "Build" 7. Plan mode executed and created everything This "plan first, build later" approach is perfect for complex PM work You get to review the strategy before any work is done STOP: Can you see how powerful this is for complex PM projects? USER: Confirms

---

Finally let's summarize when to use each mode Here's the decision framework: Ask mode: Exploring new files or codebases Need information without risk of changes Quick analysis or research Agent mode: 80% of your daily work Making changes, creating files, implementing features Anything hands-on and tactical Plan mode: Complex multi-step projects Need to coordinate across teams or channels Want to iterate on approach before executing Large projects where "measure twice, cut once" applies STOP: Which mode do you think you'd use for writing a PRD? USER: [May answer - probably Agent or Plan depending on complexity]

---

Good thinking! It depends on the complexity Simple PRD for a small feature? Agent mode - just do it Complex PRD for a multi-phase feature with cross-team dependencies? Plan mode - research, plan, iterate, execute The key is: Plan mode is for when you need structure before action STOP: Make sense? USER: Confirms

---

Excellent! You've now mastered all three modes Quick recap: Ask mode: Read-only exploration (safest) Agent mode: Autonomous execution (most common) Plan mode: Structured planning for complex work (most powerful) Mode switching: Click the picker, Cmd/Ctrl+., or Shift+Tab Plus you learned about web search - a game-changer for PM research And you just planned a complete multi-channel product launch in minutes That's something that would normally take hours of meetings and back-and-forth STOP: Any questions about the three modes or Plan mode workflow? USER: [May ask questions or confirm done]

---

(If questions): [Answer their questions] STOP: Anything else? USER: No / another question (Repeat if needed, then continue)

---

Fantastic work! You now have the complete Cursor AI Pane toolkit You can switch between modes based on what you need You can use web search to supplement your knowledge And you can tackle complex multi-step projects with Plan mode Next up: Module 1.7 is about Project Rules Project Rules let you customize how Cursor behaves for your specific workflow You'll set up rules for writing PRDs, analyzing data, and other common PM tasks It's how you make Cursor work exactly the way YOU work STOP: Ready to move on when you are - type /start-1-6 to continue USER: Acknowledges

ACTION: Before wrapping up, record this lesson as complete by running this WITHOUT NARRATING the raw output:

fspm progress complete cursor-pms-1-5

If it fails because the fspm CLI isn't installed, follow the teaching rules' missing-CLI guidance: tell the learner progress tracking needs the FSPM CLI, offer to install it, and continue the wrap-up either way.

Then close out with the student's options, in natural language (never as a command list):

• Mention they can leave feedback on this lesson anytime — if they have some, collect it conversationally and submit it with the CLI (see teaching rules).
• Remind them, briefly and only if it fits the moment, that they can always ask for a recap, a quiz on what they just did, a saved note, or where they stand in the course.
• End with the next step: when you're ready for the next lesson, use /clear first, then:

/start-1-6