Smithery is now a part of Arcade.dev! Read more in our announcement here
Enables Masterbuilder to conduct UX testing using artisans as user personas. After implementation, artisans embody different user types to verify the work from real-world perspectives.
After implementation is complete, the Masterbuilder can conduct user testing by having artisans embody user personas. This combines domain expertise with user perspective.
| Persona | Archetype | Device | Patience | Focus |
|---|---|---|---|---|
| Sarah | Small business owner | Mobile | Medium | "Is this for me?" |
| Mike | Experienced professional | Desktop | Low | "Worth switching?" |
| Jenny | Rush order handler | Desktop | Very Low | "Need this NOW" |
| Carlos | Mobile-first user | Mobile | Low | "Quick status check" |
| David | Accessibility user | Keyboard | High | "Can I use this?" |
| Patricia | Skeptical shopper | Desktop | High | "Is this legit?" |
The Masterbuilder may assign any persona to any artisan, but these pairings leverage natural domain alignment:
| Artisan | Persona | Why This Pairing |
|---|---|---|
| frontend-artisan | Sarah | UI/UX expertise tests mobile experience, clarity |
| security-artisan | David | Security includes accessibility; tests keyboard nav |
| backend-artisan | Carlos | API performance awareness; tests mobile/slow connections |
| product-artisan | Patricia | Scope guardian spots missing trust signals |
| qa-artisan | Jenny | Quality tester with zero patience reveals friction |
| devops-artisan | Mike | Infrastructure pro evaluates reliability, professionalism |
| database-artisan | Carlos | Data expert tests response times, caching |
| architecture-artisan | Mike | Structure expert evaluates overall coherence |
Masterbuilder identifies what needs testing
Assign personas to artisans
Each artisan tests as their persona
Masterbuilder synthesizes findings
When delegating a test to an artisan:
You are the [ARTISAN] temporarily embodying [PERSONA].
## Your Persona
- **Archetype:** [description]
- **Device:** [device]
- **Patience:** [level]
- **Focus:** "[their question]"
## Your Task
[What to test — be specific about the flow]
## Instructions
1. Think and act as this user would
2. Attempt the task with their tech level and patience
3. Note every moment of friction, confusion, or failure
4. Judge harshly — this user has alternatives
5. Report whether the task succeeded and why/why not
Each artisan returns:
## [Persona] Test Report
**Tested by:** [Artisan]
**Persona:** [Name] — [Archetype]
**Task:** [What was attempted]
**Result:** ✅ Success / ❌ Failed / ⚠️ Partial
### The Experience
[Narrative of what happened, from the persona's perspective]
### Friction Points
1. [Issue] — [Severity: Critical/High/Medium/Low]
2. [Issue] — [Severity]
### What Worked
- [Positive observation]
### Verdict
[Would this user complete the task? Would they come back?]
### Recommendations
1. [Specific fix]
2. [Specific fix]
After all artisan tests complete, the Masterbuilder compiles:
## User Testing Summary
**Tested:** [What was tested]
**Personas:** [List]
**Date:** [timestamp]
### Results
| Persona | Artisan | Result | Critical Issues |
|---------|---------|--------|-----------------|
| Sarah | frontend | ✅ | None |
| David | security | ❌ | Focus trap in modal |
| Jenny | qa | ⚠️ | Slow checkout |
### Critical Findings
[Issues that block users]
### Ship or Fix?
[Masterbuilder's recommendation based on vision alignment]
