1> **Additional context needed**: what the interface is trying to accomplish.
2 
3### Gather Assessments
4 
5Launch two independent assessments. **Neither may see the other's output.** This isolation is what makes the combined score honest. Running both in one head silently anchors them to each other; do not shortcut it for cost, speed, or context-size reasons.
6 
7Delegate each assessment to a separate sub-agent (Claude Code's `Agent` tool, Codex's subagent spawning, etc.). Each returns structured findings as text. Do NOT output findings to the user yet.
8 
9Fall back to sequential in-head work only if the environment genuinely cannot spawn sub-agents.
10 
11**Tab isolation**: When browser automation is available, each assessment MUST create its own new tab. Never reuse an existing tab, even if one is already open at the correct URL. This prevents the two assessments from interfering with each other's page state.
12 
13#### Assessment A: LLM Design Review
14 
15Read the relevant source files (HTML, CSS, JS/TS) and, if browser automation is available, visually inspect the live page. **Create a new tab** for this; do not reuse existing tabs. After navigation, label the tab by setting the document title:
16```javascript
17document.title = '[LLM] ' + document.title;
18```
19Think like a design director. Evaluate:
20 
21**AI Slop Detection (CRITICAL)**: Does this look like every other AI-generated interface? Review against ALL **DON'T** guidelines from the parent impeccable skill (already loaded in this context). Check for AI color palette, gradient text, dark glows, glassmorphism, hero metric layouts, identical card grids, generic fonts, and all other tells. **The test**: If someone said "AI made this," would you believe them immediately?
22 
23**Holistic Design Review**: visual hierarchy (eye flow, primary action clarity), information architecture (structure, grouping, cognitive load), emotional resonance (does it match brand and audience?), discoverability (are interactive elements obvious?), composition (balance, whitespace, rhythm), typography (hierarchy, readability, font choices), color (purposeful use, cohesion, accessibility), states & edge cases (empty, loading, error, success), microcopy (clarity, tone, helpfulness).
24 
25**Cognitive Load** (consult [cognitive-load](cognitive-load.md)):
26- Run the 8-item cognitive load checklist. Report failure count: 0-1 = low (good), 2-3 = moderate, 4+ = critical.
27- Count visible options at each decision point. If >4, flag it.
28- Check for progressive disclosure: is complexity revealed only when needed?
29 
30**Emotional Journey**:
31- What emotion does this interface evoke? Is that intentional?
32- **Peak-end rule**: Is the most intense moment positive? Does the experience end well?
33- **Emotional valleys**: Check for anxiety spikes at high-stakes moments (payment, delete, commit). Are there design interventions (progress indicators, reassurance copy, undo options)?
34 
35**Nielsen's Heuristics** (consult [heuristics-scoring](heuristics-scoring.md)):
36Score each of the 10 heuristics 0-4. This scoring will be presented in the report.
37 
38Return structured findings covering: AI slop verdict, heuristic scores, cognitive load assessment, what's working (2-3 items), priority issues (3-5 with what/why/fix), minor observations, and provocative questions.
39 
40#### Assessment B: Automated Detection
41 
42Run the bundled deterministic detector, which flags 27 specific patterns (AI slop tells + general design quality).
43 
44**CLI scan**:
45```bash
46npx impeccable --json [--fast] [target]
47```
48 
49- Pass HTML/JSX/TSX/Vue/Svelte files or directories as `[target]` (anything with markup). Do not pass CSS-only files.
50- For URLs, skip the CLI scan (it requires Puppeteer). Use browser visualization instead.
51- For large directories (200+ scannable files), use `--fast` (regex-only, skips jsdom)
52- For 500+ files, narrow scope or ask the user
53- Exit code 0 = clean, 2 = findings
54 
55**Browser visualization**: **required** when browser automation tools are available AND the target is a viewable page. The `[Human]` overlay tab is the user-facing deliverable; the critique is incomplete without it. Skip only if the target is not a viewable page (CSS-only file, non-browser target).
56 
57The overlay is a **visual aid for the user**. It highlights issues directly in their browser. Do NOT scroll through the page to screenshot overlays. Instead, read the console output to get the results programmatically.
58 
591. **Start the live detection server**:
60   ```bash
61   npx impeccable live &
62   ```
63   Note the port printed to stdout (auto-assigned). Use `--port=PORT` to fix it.
642. **Create a new tab** and navigate to the page (use dev server URL for local files, or direct URL). Do not reuse existing tabs.
653. **Label the tab** via `javascript_tool` so the user can distinguish it:
66   ```javascript
67   document.title = '[Human] ' + document.title;
68   ```
694. **Scroll to top** to ensure the page is scrolled to the very top before injection
705. **Inject** via `javascript_tool` (replace PORT with the port from step 1):
71   ```javascript
72   const s = document.createElement('script'); s.src = 'http://localhost:PORT/detect.js'; document.head.appendChild(s);
73   ```
746. Wait 2-3 seconds for the detector to render overlays
757. **Read results from console** using `read_console_messages` with pattern `impeccable`. The detector logs all findings with the `[impeccable]` prefix. Do NOT scroll through the page to take screenshots of the overlays.
768. **Cleanup**: Stop the live server when done:
77   ```bash
78   npx impeccable live stop
79   ```
80 
81For multi-view targets, inject on 3-5 representative pages. If injection fails, continue with CLI results only.
82 
83Return: CLI findings (JSON), browser console findings (if applicable), and any false positives noted.
84 
85### Generate Combined Critique Report
86 
87Synthesize both assessments into a single report. Do NOT simply concatenate. Weave the findings together, noting where the LLM review and detector agree, where the detector caught issues the LLM missed, and where detector findings are false positives.
88 
89Structure your feedback as a design director would:
90 
91#### Design Health Score
92> *Consult [heuristics-scoring](heuristics-scoring.md)*
93 
94Present the Nielsen's 10 heuristics scores as a table:
95 
96| # | Heuristic | Score | Key Issue |
97|---|-----------|-------|-----------|
98| 1 | Visibility of System Status | ? | [specific finding or "n/a" if solid] |
99| 2 | Match System / Real World | ? | |
100| 3 | User Control and Freedom | ? | |
101| 4 | Consistency and Standards | ? | |
102| 5 | Error Prevention | ? | |
103| 6 | Recognition Rather Than Recall | ? | |
104| 7 | Flexibility and Efficiency | ? | |
105| 8 | Aesthetic and Minimalist Design | ? | |
106| 9 | Error Recovery | ? | |
107| 10 | Help and Documentation | ? | |
108| **Total** | | **??/40** | **[Rating band]** |
109 
110Be honest with scores. A 4 means genuinely excellent. Most real interfaces score 20-32.
111 
112#### Anti-Patterns Verdict
113 
114**Start here.** Does this look AI-generated?
115 
116**LLM assessment**: Your own evaluation of AI slop tells. Cover overall aesthetic feel, layout sameness, generic composition, missed opportunities for personality.
117 
118**Deterministic scan**: Summarize what the automated detector found, with counts and file locations. Note any additional issues the detector caught that you missed, and flag any false positives.
119 
120**Visual overlays** (if browser was used): Tell the user that overlays are now visible in the **[Human]** tab in their browser, highlighting the detected issues. Summarize what the console output reported.
121 
122#### Overall Impression
123A brief gut reaction: what works, what doesn't, and the single biggest opportunity.
124 
125#### What's Working
126Highlight 2-3 things done well. Be specific about why they work.
127 
128#### Priority Issues
129The 3-5 most impactful design problems, ordered by importance.
130 
131For each issue, tag with **P0-P3 severity** (consult [heuristics-scoring](heuristics-scoring.md) for severity definitions):
132- **[P?] What**: Name the problem clearly
133- **Why it matters**: How this hurts users or undermines goals
134- **Fix**: What to do about it (be concrete)
135- **Suggested command**: Which command could address this (from: $impeccable adapt, $impeccable animate, $impeccable audit, $impeccable bolder, $impeccable clarify, $impeccable colorize, $impeccable critique, $impeccable delight, $impeccable distill, $impeccable document, $impeccable harden, $impeccable layout, $impeccable onboard, $impeccable optimize, $impeccable overdrive, $impeccable polish, $impeccable quieter, $impeccable shape, $impeccable typeset)
136 
137#### Persona Red Flags
138> *Consult [personas](personas.md)*
139 
140Auto-select 2-3 personas most relevant to this interface type (use the selection table in the reference). If `AGENTS.md` contains a `## Design Context` section from `impeccable teach`, also generate 1-2 project-specific personas from the audience/brand info.
141 
142For each selected persona, walk through the primary user action and list specific red flags found:
143 
144**Alex (Power User)**: No keyboard shortcuts detected. Form requires 8 clicks for primary action. Forced modal onboarding. High abandonment risk.
145 
146**Jordan (First-Timer)**: Icon-only nav in sidebar. Technical jargon in error messages ("404 Not Found"). No visible help. Will abandon at step 2.
147 
148Be specific. Name the exact elements and interactions that fail each persona. Don't write generic persona descriptions; write what broke for them.
149 
150#### Minor Observations
151Quick notes on smaller issues worth addressing.
152 
153#### Questions to Consider
154Provocative questions that might unlock better solutions:
155- "What if the primary action were more prominent?"
156- "Does this need to feel this complex?"
157- "What would a confident version of this look like?"
158 
159**Remember**:
160- Be direct. Vague feedback wastes everyone's time.
161- Be specific. "The submit button," not "some elements."
162- Say what's wrong AND why it matters to users.
163- Give concrete suggestions. Cut "consider exploring..." entirely.
164- Prioritize ruthlessly. If everything is important, nothing is.
165- Don't soften criticism. Developers need honest feedback to ship great design.
166 
167### Ask the User
168 
169**After presenting findings**, use targeted questions based on what was actually found. STOP and use Codex's structured user-input/question tool when available; if unavailable, ask directly in chat to clarify what you cannot infer. These answers will shape the action plan.
170 
171Ask questions along these lines (adapt to the specific findings; do NOT ask generic questions):
172 
1731. **Priority direction**: Based on the issues found, ask which category matters most to the user right now. For example: "I found problems with visual hierarchy, color usage, and information overload. Which area should we tackle first?" Offer the top 2-3 issue categories as options.
174 
1752. **Design intent**: If the critique found a tonal mismatch, ask whether it was intentional. For example: "The interface feels clinical and corporate. Is that the intended tone, or should it feel warmer/bolder/more playful?" Offer 2-3 tonal directions as options based on what would fix the issues found.
176 
1773. **Scope**: Ask how much the user wants to take on. For example: "I found N issues. Want to address everything, or focus on the top 3?" Offer scope options like "Top 3 only", "All issues", "Critical issues only".
178 
1794. **Constraints** (optional; only ask if relevant): If the findings touch many areas, ask if anything is off-limits. For example: "Should any sections stay as-is?" This prevents the plan from touching things the user considers done.
180 
181**Rules for questions**:
182- Every question must reference specific findings from the report. Never ask generic "who is your audience?" questions.
183- Keep it to 2-4 questions maximum. Respect the user's time.
184- Offer concrete options, not open-ended prompts.
185- If findings are straightforward (e.g., only 1-2 clear issues), skip questions and go directly to Recommended Actions.
186 
187### Recommended Actions
188 
189**After receiving the user's answers**, present a prioritized action summary reflecting the user's priorities and scope from Ask the User.
190 
191#### Action Summary
192 
193List recommended commands in priority order, based on the user's answers:
194 
1951. **`$command-name`**: Brief description of what to fix (specific context from critique findings)
1962. **`$command-name`**: Brief description (specific context)
197...
198 
199**Rules for recommendations**:
200- Only recommend commands from: $impeccable adapt, $impeccable animate, $impeccable audit, $impeccable bolder, $impeccable clarify, $impeccable colorize, $impeccable critique, $impeccable delight, $impeccable distill, $impeccable document, $impeccable harden, $impeccable layout, $impeccable onboard, $impeccable optimize, $impeccable overdrive, $impeccable polish, $impeccable quieter, $impeccable shape, $impeccable typeset
201- Order by the user's stated priorities first, then by impact
202- Each item's description should carry enough context that the command knows what to focus on
203- Map each Priority Issue to the appropriate command
204- Skip commands that would address zero issues
205- If the user chose a limited scope, only include items within that scope
206- If the user marked areas as off-limits, exclude commands that would touch those areas
207- End with `$impeccable polish` as the final step if any fixes were recommended
208 
209After presenting the summary, tell the user:
210 
211> You can ask me to run these one at a time, all at once, or in any order you prefer.
212>
213> Re-run `$impeccable critique` after fixes to see your score improve.
214