SKILL.md
Screenshot Teardown Skill
Marketing pages say what a competitor claims; screenshots show what they shipped. This skill reads real UI evidence — layout, copy, defaults, friction, what's promoted and what's buried — and turns it into competitive insight you can defend, because every claim points at pixels.
What This Skill Produces
- A screen-by-screen read: what each screenshot shows, what the design is optimising for, where the friction is
- Strategic inferences — pricing/packaging signals, target-user signals, maturity signals — each labelled as inference and tied to its visual evidence
- Learn / steal / avoid recommendations for your own product
Required Inputs
- The screenshots (up to ~5 per pass; more → ask which flow matters most). If none attached, ask — never tear down from memory of the product.
- Your product and angle (ask if missing): who's analysing, and for what decision (pricing? onboarding redesign? battlecard?)
Reading Method
- Anchor every claim to pixels. "Their onboarding asks for a credit card at step 1" — only if the screenshot shows it. Cite which screenshot each observation comes from.
- Read the hierarchy, not just the content. What's biggest, first, pre-selected, and colourful is what they want used; what's behind a "More" menu is what they don't. Defaults are strategy.
- Count the friction. Fields, steps, decisions, permission asks — visible effort before value is a measurable choice.
- Read the copy as positioning. Button labels, empty states, and upgrade nags reveal the audience and the monetisation pressure better than their homepage does.
- Separate the two registers strictly:
- Observed — on screen, citable
- Inferred — a reading of intent ("the pre-selected annual plan suggests LTV pressure"), always labelled
[inference]
- Mind the screenshot's limits. One user's session, one plan tier, one moment. Note what state the shots can't show (A/B variants, other tiers, mobile vs desktop).
