AI-orchestrated usability testing using Amazon Nova Act. The agent generates personas, runs tests to collect raw data, interprets responses to determine goal achievement, and generates HTML reports. Tests real user workflows (booking, checkout, posting) with safety guardrails. Use when asked to "test website usability", "run usability test", "generate usability report", "evaluate user experience", "test checkout flow", "test booking process", or "analyze website UX".
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
A direct command skips the review prompt. Inspect the source before running it.
The command stays on one line. Scroll horizontally to inspect it before copying.
Prefer a local copy? Download the files currently available to SkillsMP.
File Explorer
18 files
Showing SKILL.md
SKILL.md
Source instructions · Read-only preview
name
nova-act-usability
version
1.0.2
description
AI-orchestrated usability testing using Amazon Nova Act. The agent generates personas, runs tests to collect raw data, interprets responses to determine goal achievement, and generates HTML reports. Tests real user workflows (booking, checkout, posting) with safety guardrails. Use when asked to "test website usability", "run usability test", "generate usability report", "evaluate user experience", "test checkout flow", "test booking process", or "analyze website UX".
Reads:~/.openclaw/config/nova-act.json (your API key)
Writes:./nova_act_logs/ (trace files with screenshots), ./test_results_adaptive.json, ./nova_act_usability_report.html
What trace files contain:
Screenshots of every page visited
Full page content (HTML, text)
Browser actions and AI decisions
Recommendations:
Run tests only on non-production or test environments
Be aware traces may capture PII or sensitive data visible on tested pages
Review/delete trace files after use if they contain sensitive content
Consider running in a sandboxed environment (container/VM) for untrusted sites
Features
Agent-Driven Interpretation: The script no longer interprets responses. YOU (the agent) must:
Run the test script → collect raw data
Read JSON → interpret each raw_response
Set goal_achieved and overall_success
Generate the report
No hardcoded regex. No extra API calls. The agent doing the work is already running.
Quick Start (For AI Agents)
When a user asks to test a website, YOU (the AI agent) must complete ALL 4 phases:
Phase
What Happens
Who Does It
1. Setup
Generate personas, run test script
Agent + Script
2. Collect
Script captures raw Nova Act responses
Script
3. Interpret
Read JSON, determine goal_achieved for each step
Agent
4. Report
Generate HTML report with interpreted results
Agent
⚠️ The script does NOT interpret responses or generate the final report. You must do phases 3-4.
🎯 Recommended: AI Agent Generates Personas
You're already an AI (Claude) - use your intelligence to generate contextual personas!
import subprocess
import os
import sys
import json
import tempfile
# Step 1: Check dependenciestry:
import nova_act
print("✅ Dependencies ready")
except ImportError:
print("📦 Dependencies not installed. Please run:")
print(" pip3 install nova-act pydantic playwright")
print(" playwright install chromium")
sys.exit(1)
# Step 2: Verify Nova Act API key
config_file = os.path.expanduser("~/.openclaw/config/nova-act.json")
withopen(config_file, 'r') as f:
config = json.load(f)
if config.get('apiKey') == 'your-nova-act-api-key-here':
print(f"⚠️ Please add your Nova Act API key to {config_file}")
sys.exit(1)
# Step 3: YOU (the AI agent) generate personas# Example for https://www.pgatour.com/ (golf tournament site)
website_url = "https://www.pgatour.com/"
personas = [
{
"name": "Marcus Chen",
"archetype": "tournament_follower",
"age": 42,
"tech_proficiency": "high",
"description": "Avid golf fan who follows multiple tours and tracks player stats",
"goals": [
"Check current tournament leaderboard",
"View recent tournament results",
"Track favorite player performance"
]
},
{
"name": "Dorothy Williams",
"archetype": "casual_viewer",
"age": 68,
"tech_proficiency": "low",
"description": "Occasional golf viewer who watches major tournaments",
"goals": [
"Find when the next tournament is",
"See who won recently",
"Understand how to watch online"
]
}
]
# Step 4: Save personas and run testwith tempfile.NamedTemporaryFile(mode='w', suffix='.json', delete=False) as f:
json.dump(personas, f, indent=2)
personas_file = f.name
skill_dir = os.path.expanduser("~/.openclaw/skills/nova-act-usability")
test_script = os.path.join(skill_dir, "scripts", "run_adaptive_test.py")
# Run with AI-generated personas
subprocess.run([sys.executable, test_script, website_url, personas_file])
# Clean up temp file
os.unlink(personas_file)
Persona Template:
{"name":"FirstName LastName","archetype":"descriptive_identifier","age":30,"tech_proficiency":"low|medium|high","description":"One sentence about who they are","goals":["First goal relevant to this website","Second goal relevant to this website","Third goal relevant to this website"]}
📝 Alternative: Simple Custom Persona
If user specifies a persona description, pass it as a string:
# User: "Test PGA Tour site as a golf enthusiast"
website_url = "https://www.pgatour.com/"
user_persona = "golf enthusiast who follows tournaments closely"
subprocess.run([sys.executable, test_script, website_url, user_persona])
# Script will parse this and create personas automatically
⚠️ Fallback: Auto-Generation (Not Recommended)
Let the script guess personas based on basic category keywords:
# Generic, less contextual personas
subprocess.run([sys.executable, test_script, website_url])
Why YOU Should Generate Personas
✅ Advantages:
Better context: You have full conversation history and domain knowledge
Smarter inference: You can analyze the URL, industry, and user intent
No duplicate API calls: You're already Claude - don't call yourself again!
User preferences: You can adapt based on stated preferences
Clarifying questions: You can ask the user about target demographics
❌ What to avoid:
Don't let Python script make its own Claude API call (wasteful)
Don't rely on generic fallback personas (less accurate)
Don't skip persona generation (hurts test quality)
NEW: "Test the checkout process on [e-commerce site]"
NEW: "Test posting workflow on [social media site]"
The AI will automatically:
Load the Nova Act cookbook for guidance
Analyze the page to understand it
Detect if it's a workflow-based site (booking, e-commerce, social, etc.)
Generate contextual personas:
If custom persona specified → Create persona matching that description
If no custom persona → Use Claude AI to infer the 3 most plausible real-world user types
Fallback to category-based personas if AI unavailable
Create realistic test cases (including full workflows when appropriate)
Run adaptive, iterative tests with Nova Act
NEW: Apply safety stops before material impact actions (payment, posting, account creation)
Generate comprehensive HTML report with trace links
Provide viewing instructions
Workflow Testing
NEW in this version: The skill now tests complete user journeys, not just information-finding!
Supported Workflows
E-Commerce:
Product search → Add to cart → Checkout → STOP before payment
Flight/Hotel Booking:
Search → Select → Fill details → STOP before booking
Social Media:
Create post → Add content → STOP before publishing
Account Signup:
Fill registration → STOP before final submission
Form Submission:
Fill form → STOP before submit
Safety Guarantees
The skill will NEVER:
Complete actual purchases
Create real accounts
Post publicly
Send emails/messages
Subscribe to newsletters
Make any action with monetary/legal/reputational impact
The skill will ALWAYS:
Test up to (but not including) the final action
Verify the final button exists and is accessible
Document the safety stop in observations
🧠 Agent Analysis (CRITICAL)
You (the AI agent) must analyze test results! The script collects raw responses but does NOT interpret them.
Why Agent Analysis?
The script returns raw Nova Act responses like:
"No" - Is there a pricing link?
"I don't see any documentation" - Is there docs?
"Amazon Nova Act" - What is the headline?
You must determine if each response means the goal was achieved:
Response
Goal Achieved?
"No"
❌ NOT achieved
"I don't see..."
❌ NOT achieved
"Not found"
❌ NOT achieved
"Yes, I found..."
✅ Achieved
"Amazon Nova Act" (content)
✅ Achieved
"The pricing is $29/mo"
✅ Achieved
Result Data Structure
After the test script runs, read the JSON results. Each step contains:
{"step_name":"check_nav_for_pricing","prompt":"Is there a pricing link in the navigation?","expected_outcome":"Find pricing in navigation","raw_response":"No","api_success":true,"needs_agent_analysis":true,"attempts":[{"prompt":"Is there a pricing link in the navigation?","response":"No","approach":"original"}]}
Key fields you analyze:
raw_response: The actual Nova Act response - YOU determine what it means
api_success: Did the API call work? (script handles this)
needs_agent_analysis: Always true - your cue to interpret
attempts: All attempts made (script tries up to 3 alternative approaches)
How to Analyze
For each step, determine:
goal_achieved: Did the response indicate success or failure?
friction_level: How hard was it? (attempts.length > 1 = friction)
observations: UX insights from the response
Analysis example:
Step 1: "Is there a pricing link?"
→ Response: "No" (1 attempt)
→ Goal achieved: NO (explicit negative)
→ Friction: HIGH (not discoverable)
Step 2: "What is the headline?"
→ Response: "Amazon Nova Act" (1 attempt)
→ Goal achieved: YES (actual content)
→ Friction: LOW (immediately visible)
Step 3: "Find documentation"
→ Response: "I found a docs link in the footer" (3 attempts)
→ Goal achieved: YES (found eventually)
→ Friction: MEDIUM (required multiple approaches)
Helper Functions (For Script Integration)
The response_interpreter.py provides helpers if you want structured prompts:
from scripts.response_interpreter import (
format_for_agent_analysis,
create_agent_prompt_for_interpretation,
create_agent_prompt_for_alternative
)
# Format all results for analysis
formatted = format_for_agent_analysis(results)
# Get interpretation prompt for one step
prompt = create_agent_prompt_for_interpretation(step_result)
# Get retry prompt when goal not achieved
retry_prompt = create_agent_prompt_for_alternative(
original_prompt="Is there a pricing link?",
failed_response="No",
attempt_number=2
)
Complete Analysis Workflow (MANDATORY)
The script does NOT generate the final report automatically. You (the agent) must:
Run the test script → outputs test_results_adaptive.json with raw data
Read the JSON into your context
Interpret each step → set goal_achieved: true/false based on raw_response
Set overall success → set overall_success: true/false on each test
Save updated JSON
Call report generator with interpreted results
Step-by-step code for the agent to execute:
import json
import os
import sys
# Add skill scripts to path
sys.path.insert(0, os.path.expanduser("~/.openclaw/skills/nova-act-usability/scripts"))
from enhanced_report_generator import generate_enhanced_report
# 1. Read raw resultswithopen('test_results_adaptive.json', 'r') as f:
results = json.load(f)
# 2. YOU (the agent) interpret each stepfor test in results:
goals_achieved = 0for step in test.get('steps', []):
raw = step.get('raw_response', '')
# AGENT INTERPRETS: Does this response indicate goal was achieved?# You decide based on the response content and expected outcome# Example interpretations:# "No" → goal_achieved = False# "Leaderboard, News, Schedule, Players" → goal_achieved = True (content found)# "Yes" → goal_achieved = True# "I don't see any..." → goal_achieved = False
step['goal_achieved'] = ??? # YOU set this based on your interpretationif step['goal_achieved']:
goals_achieved += 1# 3. Set overall success (e.g., >= 50% goals achieved)
total = len(test.get('steps', []))
test['goals_achieved'] = goals_achieved
test['overall_success'] = (goals_achieved / total >= 0.5) if total > 0elseFalse# 4. Save interpreted resultswithopen('test_results_adaptive.json', 'w') as f:
json.dump(results, f, indent=2)
# 5. Generate report with interpreted data
page_analysis = {
'title': '...', # From your earlier analysis'purpose': '...',
'navigation': [...]
}
all_traces = []
for r in results:
all_traces.extend(r.get('trace_files', []))
report_path = generate_enhanced_report(page_analysis, results, all_traces)
print(f"Report: {report_path}")
Why the agent must interpret:
No hardcoded regex or pattern matching
You understand context (what "Yes" means for this specific question)
You can reason about partial success, edge cases
No duplicate Claude API calls - you're already running!
⚠️ Critical: Keep Nova Act Prompts Simple
Nova Act is a browser automation tool, NOT a reasoning engine.
The Claude agent (you) does all reasoning about:
What to test based on the persona
Whether results are good or bad
What the UX implications are
Nova Act just:
Clicks, types, scrolls
Reports what it sees
❌ WRONG: Asking Nova Act to reason
# DON'T ask Nova Act to think about personas
nova.act("As a beginner user, can you easily find the documentation?")
nova.act("Would a business professional find the pricing clear?")
nova.act("Is this task accomplishable for someone with low technical skills?")
✅ RIGHT: Simple, direct browser commands
# Simple browser actions
nova.act("Click the Documentation link in the navigation")
nova.act("Find and click a link containing 'Pricing'")
nova.act_get("What text is displayed in the main heading?")
nova.act_get("List the navigation menu items visible on this page")
The Correct Workflow
Agent (you) decides what to test based on persona: "Dorothy is 68 with low tech skills - she wants to know how to watch golf online"
Agent generates simple Nova Act prompts: "Click 'Watch & Listen' in the navigation"
Nova Act executes browser task and returns raw results: "Clicked Watch & Listen, now on video page"
Agent interprets results: "Dorothy would find this confusing because the options are unclear..."
How This Works
You (the AI) are the orchestrator. This skill provides:
Nova Act cookbook (references/nova-act-cookbook.md) - Best practices, workflow patterns, and safety guidelines (automatically loaded at test start)
Adaptive test orchestrator (run_adaptive_test.py) - Main execution script with workflow detection
Dynamic strategy generator (scripts/dynamic_exploration.py) - Generates workflow-appropriate test strategies
Session management (scripts/nova_session.py) - Nova Act wrapper
Report generator (enhanced_report_generator.py) - Auto-generated HTML reports
Execution Flow:
CRITICAL: Check Dependencies First
Before running ANY test, check if dependencies are installed:
# Check if nova-act is installed
python3 -c "import nova_act" 2>/dev/null
if [ $? -ne 0 ]; thenecho"Dependencies not installed. Please run:"echo" pip3 install nova-act pydantic playwright"echo" playwright install chromium"exit 1
fi# Check API keyif ! grep -q '"apiKey":.*[^"]' ~/.openclaw/config/nova-act.json; thenecho"⚠️ Please add your Nova Act API key to ~/.openclaw/config/nova-act.json"exit 1
fi
# Find the skill directory
SKILL_DIR=~/.openclaw/skills/nova-act-usability
# Run the adaptive test script
python3 "$SKILL_DIR/scripts/run_adaptive_test.py""https://example.com"# This will:# - Create nova_act_logs/ in current directory# - Create test_results_adaptive.json in current directory# - Create nova_act_usability_report.html in current directory# - Provide 60-second status updates during test
⏱️ Timeout Guidance
Recommended timeout: 30 minutes (1800 seconds)
Full usability tests with 3 personas × 3 goals = 9 tests can take 10-20+ minutes depending on:
Website load times (slow sites like media-heavy sports sites take longer)
Nova Act API response times (each act() call takes 5-60 seconds)
Network conditions
Graceful shutdown: If the test is interrupted (timeout, SIGTERM, SIGINT), it will:
Save all completed test results to test_results_adaptive.json
Generate a partial report clearly marked as incomplete
Show how many tests completed vs planned
For shorter tests: Use fewer personas or goals:
# Quick test with 1 persona
personas = [{"name": "Test User", "archetype": "casual", ...}]
What You (the AI) Need to Do:
Check dependencies (run the check above)
If missing: Tell user to run pip3 install nova-act pydantic playwright && playwright install chromium
If present: Extract the website URL from user's request
Run the test with the URL as argument
Monitor progress (status updates every 60 seconds)
Share the report viewing instructions with user
Quick Start
When user requests usability testing:
import subprocess
import os
# Get skill directory
skill_dir = os.path.expanduser("~/.openclaw/skills/nova-act-usability")
ifnot os.path.exists(skill_dir):
# Try workspace location
skill_dir = os.path.join(os.getcwd(), "nova-act-usability")
script_path = os.path.join(skill_dir, "scripts", "run_adaptive_test.py")
# Run test
result = subprocess.run(
["python3", script_path, "https://example.com"],
env={**os.environ, "NOVA_ACT_SKIP_PLAYWRIGHT_INSTALL": "1"},
capture_output=True,
text=True
)
print(result.stdout)
Detailed Workflow (Internal)
The adaptive test script (run_adaptive_test.py) handles:
Step 1: Page Analysis
Loads the page with Nova Act
Extracts title, navigation, purpose
Identifies key elements (docs, demo, pricing)
Step 2: Contextual Persona Generation
Creates personas based on what the page offers
Developer persona if API/code focused
Business persona for evaluation
Beginner persona if demo available
Step 3: Realistic Test Case Generation
Top 3 use cases per persona
Based on actual page content
Matched to persona goals
Step 4: Iterative Test Execution
For each persona + task combination:
from scripts.nova_session import nova_session
from nova_act import BOOL_SCHEMA
import time
observations = []
with nova_session(website_url) as nova:
start_time = time.time()
# Initial navigation
observations.append({
"step": "navigate",
"action": f"Loaded {website_url}",
"success": True,
"notes": "Initial page load"
})
# Execute task step-by-step (AI-orchestrated)# Break into small act() calls based on cookbook guidance# Example: "Find pricing information" task# Step 1: Look for pricing link
nova.act("Look for a link or button for pricing, plans, or subscription")
found = nova.act_get(
"Is there a visible pricing or plans link?",
schema=BOOL_SCHEMA
)
observations.append({
"step": "find_pricing_link",
"action": "Search for pricing navigation",
"success": found.parsed_response,
"notes": "Easy to find"if found.parsed_response else"Not immediately visible - UX friction"
})
if found.parsed_response:
# Step 2: Navigate to pricing
nova.act("Click on the pricing or plans link")
# Step 3: Analyze pricing page
is_clear = nova.act_get(
"Is the pricing information clearly displayed with prices and features?",
schema=BOOL_SCHEMA
)
observations.append({
"step": "view_pricing",
"action": "Accessed pricing page",
"success": is_clear.parsed_response,
"notes": "Clear pricing display"if is_clear.parsed_response else"Pricing unclear or confusing"
})
else:
# Alternative path - try search
nova.act("Look for a search function")
# ... continue orchestrating
duration = time.time() - start_time
# Document overall task result
task_success = all(obs["success"] for obs in observations if obs["success"] isnotNone)
results.append({
"persona": persona_name,
"task": task_description,
"success": task_success,
"duration": duration,
"observations": observations,
"friction_points": [obs for obs in observations ifnot obs.get("success")]
})
Step 5: Pool and Analyze Results
After all tests:
Identify common friction points across personas
Note accessibility issues for low-tech personas
Flag efficiency problems (too many steps)
Document task failures (major UX issues)
Step 6: Generate Report
import json
from scripts.enhanced_report_generator import generate_enhanced_report
# Save resultswithopen("test_results_adaptive.json", "w") as f:
json.dump(results, f, indent=2)
# Generate HTML report
report_path = generate_enhanced_report(
page_analysis=page_analysis,
results=test_results
)
print(f"Report: {report_path}")
Key Principles
Dynamic Task Decomposition
The AI should decide how to break down each task based on:
Website complexity
Persona's technical level
Task nature (navigation vs data entry vs search)
Low-tech persona example:
# More explicit, step-by-step
nova.act("Look for a button labeled 'Contact' or 'Contact Us'")
nova.act("Click on the Contact button")
result = nova.act_get("Is there a phone number or email address visible?")
High-tech persona example:
# Test efficiency features
nova.act("Look for keyboard shortcuts or quick access features")
nova.act("Try to use search (Ctrl+K or Cmd+K)")
Real-Time Observation
After EVERY act() call, analyze:
Did it succeed?
Was the UI element easy to find?
Was the label clear?
How many attempts needed?
Any error messages?
Document friction immediately in observations.
Persona-Aware Prompting
Adapt act() prompts to persona characteristics:
Elderly/low-tech: Look for obvious, labeled buttons; read everything
Power users: Try keyboard shortcuts, advanced features
Mobile users: Test mobile responsiveness, touch targets
Screen reader users: Test keyboard navigation, ARIA labels
Resources
references/nova-act-cookbook.md
MUST READ before starting any test. Contains best practices for:
Effective act() prompting
Task decomposition strategies
Data extraction patterns
Error handling
Persona adaptation
references/persona-examples.md
Template personas with detailed profiles:
Tech-savvy millennial
Elderly first-timer
Busy professional
Student/budget-conscious
Accessibility-focused
International/non-native speaker
scripts/nova_session.py
Thin wrapper providing Nova Act session primitive:
with nova_session(url, headless=True, logs_dir="./logs") as nova:
nova.act("action")
result = nova.act_get("query", schema=Schema)
scripts/enhanced_report_generator.py
Compiles observations into HTML usability report with trace file links.
assets/report-template.html
Professional HTML template for usability reports.
⚠️ IMPORTANT: First-Time Setup Required
This skill requires dependencies that must be installed before use.
For AI Agents: Dependency Check
ALWAYS check if dependencies are installed before running tests:
# Quick dependency checktry:
import nova_act
print("✅ Dependencies installed")
except ImportError:
print("📦 Dependencies not installed. Please run:")
print(" pip3 install nova-act pydantic playwright")
print(" playwright install chromium")
print("")
print("This will take 2-3 minutes to download browsers (~300MB)")
