Rate and improve a product's UX on the 11-star scale — from 'it works' (1-star) to 'it's magic' (11-star). Use when the user says 'rate the UX', '11 star', 'star rating', 'how good is the experience', 'UX audit', 'experience rating', 'what star are we at', 'how do we get to N star', or wants to evaluate and level up a product's user experience.
Rate where a product sits on the 11-star scale, then identify the smallest changes to reach the next level. Inspired by Brian Chesky's 11-star framework, adapted for software products used by both humans and AI agents.
| Stars | Name | What it means |
|---|---|---|
| 1-3 | It works | Core functionality exists. Users can accomplish the task, but it's manual, clunky, or requires documentation |
| 5 | It's useful | Organized, reliable, has the features people actually need. Users choose to use it |
| 7 | It's seamless | Zero-config where possible. Context flows automatically. Things "just work" without setup or extra steps |
| 9 | It's invisible | The tool disappears — users think about their goal, not the tool. Proactive, smart defaults, anticipates needs |
| 11 | It's magic | Users forget the tool exists because everything flows. Cross-context continuity, anticipation, genuine intelligence |
Read the project's key artifacts — README, CLAUDE.md, TASKS.md, BRIEF docs, user-facing code. Then rate each dimension:
## 11-Star Assessment: [project name]
### Current rating: [N] star
### Dimensions
| Dimension | Stars | Evidence |
|-----------|-------|----------|
| **Setup** | ? | How hard is first-time setup? |
| **Core workflow** | ? | How smooth is the main use case? |
| **Error handling** | ? | What happens when things go wrong? |
| **Discovery** | ? | Can users find features without docs? |
| **Context** | ? | Does the tool know what you need? |
| **Cross-session** | ? | Does state persist intelligently? |
| **Multi-user/agent** | ? | Does it work for teams/multi-agent? |
Overall: [weighted average or gut feel — explain which]
For each dimension, describe what the next star level looks like:
### Gap: [dimension] — [current] → [target] star
**Current:** What it does now
**Next level:** What [target]-star would look like
**Smallest change:** The minimum work to get there
**Blocked by:** Any dependencies or prerequisites
Not all dimensions matter equally. Prioritize by:
Output a ranked list of improvements, each as a potential mxit task.
Write a concrete path from current to target star level:
## Path: [current] → [target] star
### Must-haves (to claim [target] star)
- [ ] Change 1 — why it matters
- [ ] Change 2 — why it matters
### Nice-to-haves (solidifies the rating)
- [ ] Change 3
- [ ] Change 4
### What NOT to build (belongs at a higher star level)
- Thing X — this is [higher] star territory, premature now
11star assessment is a natural input to the fold loop:
Run 11star at the start of an iteration (where are we?) and at the end (did we level up?).
TigerFlare used this framework to go from 5-star to 7-star:
deno task start does everything. MCP provides context on connect. Save skill works cross-project. 7-star achieved.See BRIEF-11STAR.md in TigerFlare for the full assessment.