Skip to content
bettersense

Tutorial: Set up your profile

Time: ~10 minutes
Skills used: user-profile
Who this is for: Everyone. This is the first thing to do after installing bettersense.

You’ll learn:

  • What the profile interview actually asks and why each question matters
  • What a finished profile looks like
  • Which skills read it automatically — and what changes once it’s there

Prerequisites: bettersense installed. That’s it.

Why this first

Section titled “Why this first”

Most skills in the bundle produce generic output by default. They don’t know whether you’re a PM or an EM, whether you’re at a startup or an enterprise, whether you prefer direct feedback or diplomatic framing. You end up re-establishing that context every session.

user-profile fixes this once. After a 10-minute interview, the result is a single private file — ~/bettersense-work-reflections/profile.md — that 8+ skills read automatically. You explain yourself once; the skills adjust forever.

It also creates ~/bettersense-work-reflections/, the directory where all your bettersense data lives, and shows you a one-time privacy notice before writing anything.

The scenario

Section titled “The scenario”

You’re a senior PM at a mid-size B2B SaaS company. You’ve just installed bettersense. This tutorial follows what your first session looks like.

Step 1: Start the interview

Section titled “Step 1: Start the interview”

In Claude Code, type:

set up my profile

user-profile fires. Before asking anything, it shows you the privacy notice:

“I’m about to create ~/bettersense-work-reflections/, which will hold your profile and any reflections you log over time. This directory lives on your local machine only — not synced, not shared. Confirm the location, or tell me a different path if you’d prefer (e.g. an encrypted volume).”

You confirm. The interview begins.

Step 2: The interview (what it asks and why)

Section titled “Step 2: The interview (what it asks and why)”

The skill walks through six questions. None are required — you can skip any — but each one pays off in a specific set of skills.

1. Role and level

“Start with your role. Title, level, how long you’ve been in it, and roughly who you report to — no names, just the shape. Something like ‘Senior PM, 3 years, report to a Director of Product who reports to the VP.’”

You answer:

“Senior PM, 2.5 years in role. I report to a Director of Product. My skip is the VP of Product. I lead a team of 3 engineers and 1 designer.”

Why it matters: report-promo-case and report-career-architect calibrate the level bar against your current title. coaching-mode and feedback-frameworks adjust their framing based on whether you’re a manager or IC.

2. Org and product context

“Two or three sentences on the company and what you own — stage, size, your product area, your primary cross-functional partners.”

“Series C, ~200 people. I own the core workflow product — the thing our customers use every day. Primary partners are engineering, design, and CS. I work closely with the growth PM on shared metrics.”

Why it matters: the-spec-writer anchors specs to your product context. metrics-design doesn’t suggest consumer-style metrics at an enterprise B2B company. the-explainer calibrates the audience register.

3. Current strategic focus

“What are the 2-4 things you’re spending most of your time on this quarter? Not a task list — the narrative arcs you’re trying to advance.”

“1. Reducing time-to-value for new customers — we have an onboarding drop-off problem. 2. Shipping the collaboration features we’ve been holding for 18 months. 3. Getting better at measuring AI feature quality — we’ve shipped two LLM-powered features without real evals.”

Why it matters: prioritization-frameworks scores backlog items against these tracks. product-pulse watches metrics tied to them. patterns-watch flags when your logged work drifts from your stated focus.

4. Communication style

“How do you want outputs framed? Direct or diplomatic? Examples or abstractions? Should the skills push back on you, or stay in support mode?”

“Direct. No softening. I want pushback when my framing is off. Examples over abstractions — I’m a concrete thinker. Keep it short.”

Why it matters: This shapes every skill’s output register. A profile that says “direct, no softening” produces a different feedback-frameworks output than one that says “diplomatic, my team is sensitive to criticism.”

5. What you’re working on as a leader

“One or two things you’re actively trying to get better at as a leader. Optional, but high value — this composes with coaching, feedback, and self-reflection.”

“I tend to solve problems for my team instead of coaching them through it. I’m working on asking more questions before giving answers.”

Why it matters: coaching-mode will reference this when you’re in a development conversation — it knows what you’re trying to practice. self-reflect sessions accumulate against this. patterns-watch will surface whether your logged behavior matches what you say you’re working on.

6. Stack and tools (optional)

“Only if relevant — languages, frameworks, observability stack. Useful for the-spec-writer, the-eval-designer, the-rfc-reviewer, the-prompt-critic.”

“Python backend, React frontend. We use PostHog for product analytics, Datadog for infra monitoring. AI stack is Claude via the Anthropic API.”

Step 3: Review and confirm

Section titled “Step 3: Review and confirm”

The skill drafts the profile and shows it to you before writing:

“Here’s what I’ll write to ~/bettersense-work-reflections/profile.md. Anything to change?”

---
name: (optional)
last_updated: 2026-05-08
default_hat: pm
---


## Role
Senior PM, 2.5 years in role. Reports to a Director of Product who reports to the VP of Product.
Leads a team of 3 engineers and 1 designer.


## Org and product context
Series C, ~200 people. Owns the core workflow product. Primary cross-functional partners:
engineering, design, CS, and growth PM on shared metrics.


## Current strategic focus
- Reducing time-to-value for new customers (onboarding drop-off)
- Shipping collaboration features (18 months in the backlog)
- Building real eval infrastructure for AI features


## Communication style
Direct. No softening. Pushback welcome when framing is off. Examples over abstractions. Keep it short.


## What I'm working on as a leader
Tendency to solve problems for the team rather than coaching. Practicing asking questions
before giving answers.


## Stack and tools
Python backend, React frontend. PostHog (product analytics), Datadog (infra).
Claude via Anthropic API for AI features.

You confirm. File written.

What changes now

Section titled “What changes now”

From this session forward, every skill that reads profile.md starts with your context already loaded:

  • the-spec-writer — specs anchor to your product area and strategic tracks automatically; no more explaining what you own
  • the-translator — exec comms are framed for your org size and audience register
  • report-promo-case — the promotion bar is calibrated to your current level
  • coaching-mode — knows you’re trying to ask more questions; will reinforce that rather than letting you prescribe
  • feedback-frameworks — drafts feedback in your direct register, not a softened default
  • metrics-design — suggests B2B SaaS metrics, not consumer growth metrics
  • patterns-watch — watches whether your logged work matches the strategic focus you named

Skills that don’t read the profile still work — outputs are just less calibrated.

Keeping it current

Section titled “Keeping it current”

The profile is a semi-static anchor. Update it when something material changes: a new role, a significant scope shift, a new strategic focus at the start of a planning cycle. Don’t update it every week.

After 90 days, user-profile will gently ask whether anything has changed when you invoke it. It won’t nag in between.

update my profile

The skill will show you the current file and let you edit specific sections rather than starting over.

Where to go next

Section titled “Where to go next”