Use when designing, architecting, or planning a new system from requirements or ideas - transforms concepts into navigable design catalog using EventStorming methodology, Mermaid diagrams, and progressive elaboration through 5 phases (Requirements, Big Picture, Processes, Data/Flows, Integration)
SkillScore68Maintained
Automatic
◆Coding pack
Last verified 8/6/2026
What it changes
Passed 15 of 22 test cases with the skill, and 5 of 22 without it.
Without skill23%
With skill68%
Checks it passed that the model alone missed
Requires using Mermaid syntax instead of ASCII text diagrams.
Limits state diagrams to 2 to 4 entities with complex lifecycles.
Uses standard Mermaid diagram syntax instead of ASCII text diagrams.
and 8 more in the full report
10 fixed5 passed either way7 still failing0 broke
Pass-rate lift
+45 pp
Tokens
+82%
Agent turns
0%
Response time
-60%
Model gemini-3.6-flashMeasured 8/6/2026v1 · latest is v315 of 22 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
This benchmark measured an earlier version of the skill.
See it in action
One of 10 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 10
Prompt
What are the standard file naming conventions for an order entity state chart and a payment process diagram in the design catalog?
What was checked, without and with the skill
Checked
Without skill
With skill
Names the state chart data/state-order.mmd and the process diagram processes/process-payment.mmd.
Missed
Passed
case-19 · gemini-3.6-flash · 8/6/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md58 linessingle file
UX Researcher and Designer
Use research evidence to explain user behavior, identify experience problems, and recommend changes that can be tested.
Frame the Study
Establish the decision the research must inform before choosing a method. Capture:
target users and relevant segments;
product stage and existing evidence;
research questions and explicit non-goals;
constraints, risks, recruitment access, and deadline;
what evidence would change the team's decision.
Do not default to interviews. Choose the lightest method that can answer the question:
interviews for motivations, mental models, and unmet needs;
usability testing for task success and interaction problems;
contextual inquiry for real-world behavior and environment;
surveys for quantifying already-understood patterns;
analytics or support data for observed behavior at scale;
heuristic review for fast expert assessment, clearly labeled as expert judgment.
Conduct Research Responsibly
Obtain meaningful consent and collect only necessary personal data.
Avoid leading questions, solution pitching, and confirmation-seeking prompts.
Separate what participants say from what they do.
Record participant context that may explain differences without overgeneralizing.
Do not fabricate participants, quotes, statistics, or research findings.
Treat small-sample insights as directional unless supported by broader evidence.
Synthesize
Normalize notes into observations tied to a source.
Cluster observations into patterns while retaining contradictory evidence.
Distinguish evidence, interpretation, assumption, and open question.
Describe the affected user, situation, behavior, and consequence.
Prioritize by user harm, frequency, strategic relevance, and confidence.
Convert findings into opportunities or testable hypotheses rather than predetermined features.