comparable-apps
Comparable Apps
Related skills: use one-sentence-test to write the cohesion sentence that the comparable-apps check audits against, and cohesive-clean-breaks when the comparison surfaces an asymmetric win (refuse a feature).
Core move. Before designing a user-facing surface, name 3-5 comparable apps and write one line each about what they do for the same problem. Then ask: which category is Epicenter in, and does the proposed design match that category's pattern? When the design diverges from the category, name the reason. When it converges, the comparison is your evidence that the choice is unsurprising.
Show your work. Write the table out in the spec or design doc, visibly. The value is in the reader seeing the pattern (or the deviation from the pattern) explicitly, not hidden inside your reasoning.
Why this skill exists
Epicenter is a local-first workspace platform. Many design questions ("should we cache email", "should we show identity in chrome", "should the sync state be visible", "do we need multi-account") have well-trodden answers in comparable categories. Asking the question explicitly catches two failure modes:
- Importing a pattern from the wrong category. Caching every profile field because Gmail does is wrong; Gmail is communication-first and Epicenter is not.
- Inventing a novel pattern when a settled one exists. Designing a new identity surface when Linear/Notion/Cursor already converged on "avatar in chrome, email in account menu" is gratuitous.
The skill is a coherence gate, not a recipe. Categories are heuristic. When the comparison says "do X" but Epicenter has a load-bearing reason to do Y, the reason goes in the spec; the comparison surfaced the cost of the deviation.