AI Agent Platform for Temporal Code Intelligence
The missing evolution layer that AI code agents need
git clone https://github.com/Cre4T3Tiv3/gitvoyant.git
cd gitvoyant
make bootstrap
๐ Complete Setup Guide & Documentation โ
GitVoyant v0.2.0 is built using Domain-Driven Design (DDD) principles with Clean Architecture patterns, ensuring maintainable, extensible, and reliable code:
- ๐ฏ Domain Layer: Rich entities (
TemporalEvaluation
) and value objects (ComplexityTrend
,ConfidenceRank
) with embedded business logic - ๐๏ธ Application Layer: Use cases orchestrating business workflows with clear separation of concerns
- ๐ก Infrastructure Layer: Core temporal intelligence engine with Git integration and statistical analysis
- ๐จ Presentation Layer: CLI interface and AI agent integration with natural language capabilities
Architecture Philosophy: Clean boundaries between domain logic, infrastructure, and presentation layers enable independent evolution and testing of each component.
Current Test Coverage: 55+%
- โ Unit Tests: Domain entities, value objects, and core algorithms
- โ Integration Tests: Full workflow testing with real Git repositories
- โ Agent Tests: AI agent interaction and tool integration
- โ CLI Tests: Command-line interface and intuitive output formatting
Comprehensive test suite ensures reliability across temporal analysis, Git integration, and AI agent workflows.
Live Temporal Intelligence: The Claude-powered agent provides conversational access to GitVoyant's temporal analysis engine:
๐ฌ You: "Which files in this repo need the most attention?"
๐ค Claude: Analyzing repository temporal patterns...
๐ฆ Repository decay risks:
src/api/handlers.py: 0.85 (HIGH RISK - complexity growing +2.3/month)
utils/data_processing.py: 0.72 (MEDIUM RISK - declining pattern detected)
core/business_logic.py: 0.68 (MEDIUM RISK - confidence: 0.4 - limited history)
Note: Currently supports Claude AI only. Multi-LLM support (GPT-4, Gemini, Local LLMs) planned for v0.3.0.
- ๐ฌ Confidence Scoring: Statistical confidence based on commit history depth
โ ๏ธ Low Confidence Warnings: Transparent reliability indicators for limited data- ๐ Progressive Disclosure: Higher confidence with more temporal data points
- ๐ฏ Improved Risk Assessment: Enhanced quality decay forecasting algorithms
- ๐๏ธ Domain-Driven Architecture: Clean separation with rich domain entities
- ๐ฎ Claude + LangChain integration for natural language interaction
- ๐ง AI Agent Runtime: Conversational temporal intelligence analysis
- ๐ Comprehensive Testing: 55+% test coverage across all layers
GitVoyant provides a Rich-formatted CLI experience:
Visual Features:
- ๐ Command Structure: Organized help system with clear options
- ๐ Auto-Completion: Shell completion support for enhanced productivity
- ๐ Typography: Clean terminal formatting with proper spacing
Deep temporal analysis with statistical confidence and visual indicators.
What You See:
- ๐ฎ Banner: Welcome message with version and branding
- ๐ Info Grid: Repository stats with ๐๐๐ indicators
- ๐ Repository path
- ๐ Health score (-0.35 = improving complexity)
- ๐ Number of evaluated files
- ๐ Results Table: Color-coded temporal scores
- ๐ข Green Circle: -0.35 indicates decreasing complexity
- File paths in cyan
- Clean table formatting
Status Messages:
- ๐ Info: "Initializing temporal analysis..." in cyan
- โ Success: "Temporal evaluation complete" in green with checkmark
Conversational temporal intelligence powered by Claude AI.
Agent Experience:
- ๐ง Natural Language: Ask questions in plain English about code quality
- ๐ฌ Interactive Dialog: Real-time conversation with temporal analysis
- ๐ Detailed Analysis: Specific decay rates, risk scores, and recommendations
Conversation Flow:
๐ฌ You: What is the decay rate of src/gitvoyant/cli/analyze.py?
๐ค Claude: Based on the analysis:
- The file shows a negative trend of -0.35 per month, indicating decreasing complexity over time
- It has LOW exposure and a risk score of 0.00
- The analysis is based on 11 commits
- Overall, this file appears to be well-maintained with improving code quality and minimal decay risk
Agent Capabilities:
- ๐ File-Specific Analysis: Deep dive into individual file patterns
- ๐ Trend Interpretation: Explains what temporal patterns mean
- ๐ก Actionable Insights: Provides specific recommendations
- ๐ Statistical Context: Includes commit counts and confidence levels
GitVoyant implements modern terminal UI principles:
- Indicators: ๐๐๐ for quick scanning
- Color Coding: Green/yellow/red for pattern recognition
- Typography: Clean spacing and alignment
- Feedback: Status messages during operations
- Completion: Shell auto-completion support
- Responsive: Adapts to terminal width
- Consistency: Unified experience across commands
- Hierarchy: Key information first
- Scannable: Tables with clear headers
- Contextual: Descriptive help and options
- Transparent: Clear status for remote operations
from gitvoyant import TemporalEvaluatorService
service = TemporalEvaluatorService()
evaluation = await service.analyze_file("src/main.py")
print(f"Quality Pattern: {evaluation.quality_pattern}")
print(f"Complexity Trend: {evaluation.complexity_tenor.slope:.2f}/month")
print(f"Confidence: {evaluation.confidence_score:.2f}")
if evaluation.confidence_warning:
print(f"โ ๏ธ {evaluation.confidence_warning}")
repo = await service.analyze_repository(".")
print(f"Health Score: {repo.overall_health_score}/10.0")
print(f"Quality Distribution: {repo.quality_distribution}")
# Identify files with quality improvement patterns
improving = repo.improving_files
print(f"Files showing improvement: {len(improving)}")
from gitvoyant.application.agent_runtime import create_gitvoyant_agent
agent = create_gitvoyant_agent()
response = agent.invoke({
"input": "Which files in this repository need attention?"
})
print(response["output"])
# Analyze a specific file
gitvoyant analyze temporal ./repo src/main.py
# Launch interactive AI agent
gitvoyant analyze agent
# Get help
gitvoyant --help
๐ Complete CLI Reference โ
GitVoyant's core algorithm implements sophisticated temporal pattern recognition:
# Extract complexity metrics across Git history
commits = repo.iter_commits(paths=file_path, since=analysis_window)
evolution_data = [extract_metrics(commit) for commit in commits]
def _cyclomatic_complexity(self, ast_tree):
complexity = 1 # Base complexity
for node in ast.walk(ast_tree):
if isinstance(node, (ast.If, ast.While, ast.For)):
complexity += 1
elif isinstance(node, ast.BoolOp):
complexity += len(node.values) - 1
return complexity
# Compute temporal trend using statistical analysis
complexity_trend = np.polyfit(range(len(data)), complexity_values, 1)[0]
quality_pattern = classify_pattern(complexity_trend) # IMPROVING/DECLINING/STABLE
# Enhanced confidence scoring based on data quality
if commit_count >= 10: confidence_score = 0.9
elif commit_count >= 7: confidence_score = 0.75
elif commit_count >= 5: confidence_score = 0.6
else: confidence_score = 0.4 # Low confidence warning
# Quality decay forecasting
quality_decay_forecast = min(max(complexity_growth_rate * 2, 0), 1)
Key Innovation: GitVoyant transforms raw commit data into actionable temporal intelligence through statistical analysis, pattern recognition, and predictive modeling.
Static AI: "This file has high complexity."
GitVoyant: "This file is reducing complexity monthly, this is quality engineering."
git clone https://github.com/Cre4T3Tiv3/gitvoyant.git
cd gitvoyant
make bootstrap
cp .env.example .env
# Edit .env with your ANTHROPIC_API_KEY
ANTHROPIC_API_KEY
for full AI agent features
- User Guide - Complete setup, usage, and CLI reference
- Contributing Guide - Development setup, guidelines, and contribution opportunities
- Complexity Requirements - Deep dive into the complexity metrics
- Temporal Analysis Explained - Deep dive into the science
- ๐ค Multi-LLM Support: GPT-4, Gemini, Claude, Local LLMs (Ollama)
- ๐ฌ Advanced Algorithms: R-squared confidence, trend stability metrics
- ๐ Multi-Language Support: JavaScript, Java, Go temporal patterns
- ๐ Enhanced Metrics: Maintainability index, technical debt scoring
- ๐ Pattern Mining: Cross-repository quality pattern discovery
- ๐ Web Dashboard: Interactive temporal visualization and insights
- ๐ CI/CD Integration: GitHub Actions, GitLab CI temporal monitoring
- ๐ฅ Team Analytics: Developer impact analysis and collaboration patterns
- ๐ Predictive Models: ML-enhanced quality decay forecasting
- ๐ฏ Custom Patterns: Organization-specific quality pattern recognition
- ๐ง Agent Ecosystem: Specialized agents for different analysis domains
- ๐ฎ Temporal Recommendations: AI-powered refactoring suggestions
- ๐ Knowledge Base: Historical pattern learning and recommendation
- ๐ค IDE Integration: VSCode, IntelliJ temporal intelligence plugins
- ๐ Community Platform: Shared pattern libraries and best practices
- โ Claude AI (Anthropic) - Full support
- ๐ง GPT-4, Gemini, Local LLMs - Planned for v0.3.0
- โ Python - Complete temporal analysis
- ๐ง JavaScript, Java, Go - Research in progress
- โ Commit-based confidence scoring
- ๐ง R-squared trend stability - Enhanced algorithms planned
GitVoyant uses temporal pattern evaluation to identify code evolution signatures:
- ๐ Temporal Complexity Analysis โ Tracks cyclomatic complexity evolution with linear regression
- ๐ฏ Quality Pattern Recognition โ Classifies IMPROVING/DECLINING/STABLE patterns with confidence scoring
- ๐ Change Pattern Analysis โ Maps modification patterns to quality outcomes
- ๐ฎ Risk Correlation Modeling โ Predicts future maintenance burden with decay forecasting
- ๐ฅ Developer Impact Assessment โ Analyzes how team changes affect code health
- ๐ Statistical Confidence โ Progressive disclosure based on temporal data quality
๐ Learn more: Temporal Analysis Explained
- โ Temporal Pattern Recognition: IMPROVING/DECLINING/STABLE classification with statistical rigor
- โ Confidence Scoring: Progressive disclosure based on temporal data quality (55+% test coverage)
- โ Statistical Analysis: Linear regression trend analysis with R-squared validation
- โ Real-World Ready: Handles real Git histories and large projects with graceful degradation
- โ Domain-Driven Design: Rich entities and value objects encoding business logic
Code Review โ Find Current Issues โ Fix Current Issues
๐ฅ Limitation: Reactive and blind to positive evolution
History Analysis โ Pattern Recognition โ Confidence Assessment โ Future Guidance
โ Advantage: Proactive insights, quality discovery, predictive power, statistical honesty
GitVoyant is open source and welcomes contributions! From AI agent enhancements to temporal intelligence research, there are many ways to contribute to the future of code analysis.
๐ Complete Contributing Guide โ - Comprehensive guidelines, research areas, and development standards
git clone https://github.com/Cre4T3Tiv3/gitvoyant.git
cd gitvoyant
make bootstrap
make dev
๐ Development Workflow Details โ
- ๐ฌ Cross-language quality pattern mining
- ๐ค Agent training via historical code context
- ๐ Predictive engineering effectiveness models
- ๐ CI-integrated code health monitoring
- ๐ Statistical confidence and trend stability
- ๐ง Multi-LLM temporal intelligence integration
GitVoyant v0.2.0 is licensed under Apache 2.0.
See LICENSE
๐ฎ GitVoyant by Jesse Moses (@Cre4T3Tiv3) at ByteStack Labs
Could temporal intelligence with statistical confidence be the missing foundation for AI-native engineering?
git clone https://github.com/Cre4T3Tiv3/gitvoyant.git
cd gitvoyant
make bootstrap
gitvoyant analyze temporal . --help
GitVoyant v0.2.0 - Where Temporal Intelligence Meets AI Engineering
Built with Domain-Driven Design โข Powered by Statistical Analysis โข Enhanced by AI