strt-where-to-start

name	strt - Where to Start
description	For when you have a goal but genuinely don't know where to begin. Find the zero-dependency starting point and take it.
output	{"format":"prose"}

STRT - Where to Start

Input: $ARGUMENTS

---

Core Principles

1. The start is the thing nothing else depends on. Every project has tasks that require other tasks to be done first. The starting point is the task that requires nothing else. Find it.

2. Overwhelm is a starting problem, not a planning problem. When everything feels like too much, you don't need a better plan. You need a smaller first action.

3. Starting wrong is better than not starting. Any starting point that teaches you something is a valid start. Perfecting your start is a form of procrastination.

4. Clarity comes from action, not thought. If you can't figure out where to start by thinking about it, start anywhere and the right path will reveal itself.

---

Step 1: What's the End State?

GOAL: [what "done" looks like — from $ARGUMENTS]

Be specific. "Launch the app" is better than "finish the project." If the user can't state the end state, the starting point is: define the end state.

Step 2: List Everything That Seems Required

REQUIRED (brain dump — not ordered, not filtered):
  1. [thing]
  2. [thing]
  3. [thing]
  ...

Don't organize. Don't prioritize. Just list every task, requirement, or thing that needs to happen. Include vague ones. Include ones you're not sure about.

Step 3: Find the Dependencies

For each item, ask: "Does this require any other item on the list to be done first?"

DEPENDENCY MAP:
  [item] ← depends on nothing
  [item] ← depends on [other item]
  [item] ← depends on [other item], [other item]
  ...

Step 4: Find the Zero-Dependency Item

ZERO-DEPENDENCY ITEMS: [items that depend on nothing else]

If there are multiple, pick the one that:

• Unblocks the most other items
• Is smallest in scope
• Gives you the most information about whether your plan is right

→ START HERE: [the one item]
  WHY: [what it unblocks or reveals]
  FIRST ACTION: [the literal first thing to do — open a file, make a call, write a sentence]

Step 5: Confirm It's Actually Doable Right Now

CAN YOU DO THIS RIGHT NOW?
  - Do you have what you need? [yes/no — if no, THAT is the real start]
  - Do you know how? [yes/no — if no, "learn how to [X]" is the real start]
  - Is anything stopping you? [yes/no — if yes, removing that blocker is the real start]

ACTUAL START: [confirmed first action]

---

Overwhelmed Start

If the user says they're overwhelmed, or the list in Step 2 has more than 10 items, or they can't complete Step 2:

OVERWHELMED START
=================
Skip everything above. Do this instead:

1. What's the SMALLEST possible action related to your goal?
   Not the most important. Not the most logical. The SMALLEST.
   (Examples: open the document, write one sentence, send one message, read one page)

2. Can you do it in under 5 minutes?
   If no, make it smaller.

3. → DO THAT NOW.

You don't need to know the whole path. You need to take one step.
After that step, you'll know more about what the next step should be.

---

Output Format

WHERE TO START
==============

GOAL: [end state]

EVERYTHING REQUIRED:
  [numbered list]

DEPENDENCIES:
  [zero-dep items marked with →]

→ START HERE: [chosen starting point]
  FIRST ACTION: [literal next physical/digital action]
  THIS UNBLOCKS: [what becomes possible after]

---

Failure Modes

Failure	Signal	Fix
Planning instead of starting	Spent >10 min on Steps 2-3	Switch to Overwhelmed Start
No zero-dependency items	Every item depends on another	You have a circular dependency. Break one arbitrarily
Start is too big	First action takes more than 1 hour	Break it down further. What's the first 15 minutes?
"I need to research first"	Research as starting point	Fine, but scope it: research WHAT, for HOW LONG, looking for WHAT ANSWER
Can't list requirements	Step 2 produces nothing	You don't understand the goal yet. Start: talk to someone who does, or write down what you don't know

---

Integration

• If user is stuck mid-project (not at the start): route to /unstk
• If user knows where to start but needs the next step: route to /wtdn or /nstep
• If the goal itself is unclear: route to /gu
• For complex project planning: route to /de or /to
• Differs from /next: next assumes you're mid-process; strt assumes you haven't begun