.cursor/skills/experiment-testing-standards/SKILL.md
--- name: experiment-testing-standards description: Enforces experiment testing standards: real implementation testing, no simplified approximations, exact formula matching. Use when creating experiments, testing algorithms, or validating mathematical implementations. --- # Experiment Testing Standards ## Core Principle **Experiments must test the REAL implementation, never simplified approximations.** ## Mandatory Requirements ### ✅ REQUIRED - **Real Implementation Logic** - Use exact same
npx skillsauth add avra-cadavra/avrai .cursor/skills/experiment-testing-standardsInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Experiments must test the REAL implementation, never simplified approximations.
# Simplified distance decay (NOT the real implementation)
location_compatibility = exp(-decay_rate * distance_km)
# Real quantum state inner product (matches production code)
location_state = create_location_quantum_state(
lat, lon, type, accessibility, vibe
)
compatibility = abs(inner_product(location_state_a, location_state_b)) ** 2
All experiments MUST use real Big Five OCEAN data:
from shared_data_model import load_and_convert_big_five_to_spots
# Load and convert Big Five OCEAN to SPOTS 12
spots_profiles = load_and_convert_big_five_to_spots(
max_profiles=1000,
data_source='auto',
project_root=project_root
)
❌ FORBIDDEN:
.cursorrules - Experiment Testing Standards sectiondocs/plans/test_refactoring/TEST_WRITING_GUIDE.mddevelopment
--- name: world-model-development description: Guides world model development patterns: state/action encoders, ONNX inference, feature extraction pipeline, latency budgets. Use when implementing world model components, state encoders, action encoders, feature extractors, or ONNX models. Core skill for Phases 3-6. --- # World Model Development Patterns ## Core Principle All world model components follow LeCun's autonomous machine intelligence framework. State observations flow through a percep
tools
Implements base workflow controller patterns for multi-step processes. Use when creating complex workflows that require orchestration of multiple steps with error handling and rollback.
testing
--- name: widget-test-patterns description: Guides widget test patterns: BLoC testing, user interactions, state changes, material app setup. Use when writing widget tests, testing UI components, or validating widget behavior. --- # Widget Test Patterns ## Core Pattern Widget tests verify UI behavior: user interactions, state changes, and visual display. ## Basic Widget Test Setup ```dart testWidgets('widget displays correctly', (WidgetTester tester) async { // Arrange: Create widget awa
testing
--- name: test-template-generation description: Generates test templates: unit, widget, integration, service tests following project patterns. Use when creating new tests or ensuring tests follow project standards. --- # Test Template Generation ## Available Templates Test templates are located in `test/templates/`: - `unit_test_template.dart` - `widget_test_template.dart` - `integration_test_template.dart` - `service_test_template.dart` ## Unit Test Template ```dart /// SPOTS Component Uni