/crisp-research
See how others solved it before you design.
(What it does)
The research layer. /crisp-research takes a feature brief, works out what type of feature it is, and searches a set list of design sources (Mobbin, SaaSpo, Screenlane, Dribbble, and written case studies) for named patterns. It links each pattern to the CRISP dimensions most at risk for that type of feature.
The output gives you direction, not a mood board: patterns worth studying, anti-patterns (common mistakes) not to copy, benchmarks to beat, and the open questions the brief left unanswered. It feeds straight into /feature-design.
(When to use it)
Use it at the start of feature work, before you open a reference site yourself:
- You are starting a feature and about to lose an afternoon browsing Mobbin and Dribbble
- A product brief has landed and you suspect gaps: states, permissions, or mobile
- You want named, sourced patterns rather than a folder of screenshots
- You want
/feature-designto work from evidence instead of taste
(What you get)
- Summary of patterns from other products
- CRISP dimension risk flags
- Anti-patterns to avoid
- Gaps in the brief, found before design starts
(Example output)
Illustrative example only. This is not a real audit.
(What it will not do)
- No raw URLs or screenshot dumps. Only named patterns and named products
- No more than three patterns per dimension, ever
- No padding when results are thin. It says its confidence is low instead
- No design recommendations. That is the job of
/feature-design - No following the crowd. Overused visual styles are flagged so the design deliberately moves away from AI-generated defaults
(Common questions)
What input does it need?
A feature name, a problem statement, or a pasted product brief. If the brief is too vague to search on, it asks one clarifying question and waits. It will not waste searches on a guess.
What is a saturated lane warning?
A warning that the brief is heading towards a visual style everyone already uses, such as the type-led editorial landing page, the generic cream SaaS dashboard, or the glowing dark AI look. The warning names the style and suggests other directions, so /feature-design makes a deliberate choice.
How do the sources differ?
Mobbin for interface patterns from shipped apps, SaaSpo for B2B and enterprise SaaS, Screenlane for micro-interactions, Dribbble for visual direction only (and labelled as such), and web search for written case studies. The order of priority changes with the feature type.
What are the open questions in the output?
The specific gaps the brief left, such as states, permissions, mobile, or triggers, raised as no more than three named questions for the product manager. Every brief has gaps. The research names the ones that change the design.
(Install)
Installs all fourteen CRISP skills and detects which AI coding tool you use. Run /crisp-teach once per project, before the other skills.
(Related skills)