predictive-analysis

name	predictive-analysis
description	Predictive code analysis for identifying potential risks, anti-patterns, and future maintenance issues. Use when assessing code health or planning refactoring.
disable-model-invocation	false
user-invocable	true
allowed-tools	Read, Grep, Glob, Bash
argument-hint	["target-path"]

Predictive Code Analysis

Analyzes codebases to proactively predict and identify risks that may become problems in the future.

Overview

The predictive code analysis system evaluates the following risk categories:

• Technical debt: Growing complexity, code that is difficult to maintain
• Security risks: Insufficient input validation, vulnerability patterns
• Performance bottlenecks: Inefficient algorithms, scalability issues

Basic Usage

# Analyze entire project
/predictive-analysis

# Analyze a specific directory
/predictive-analysis src/

# Analyze by component
/predictive-analysis src/components/

Risk Assessment Framework

Risk Levels

Level	Description	Response Deadline
Critical	Requires immediate fix, impacts production	Immediately
High	Requires urgent response, potential future failure	Within 1 week
Medium	Requires planned response, impacts maintainability	Within 1 month
Low	Improvement recommended, long-term quality	Next refactoring session

Evaluation Criteria

Each risk is evaluated from the following perspectives:

1. Likelihood: The probability that this issue will actually occur
2. Impact: The severity of damage if it occurs
3. Timeline: Predicted timeframe for the issue to surface
4. Effort: Comparison of effort to fix now vs. fix later

Analysis Process

1. Pattern Recognition

• Detection of common code patterns that cause problems
• Identification of hotspots where complexity is growing
• Discovery of anti-patterns that break down at scale
• Identification of potential time bombs (hardcoded values, assumptions)

2. Risk Classification

The analysis classifies findings into the following categories:

Technical Debt

• High-complexity functions
• Code duplication
• Tight coupling
• Frequently changed files

Security Risks

• Insufficient input validation
• Weak authentication/authorization
• Exposure of secrets
• Unsafe data processing

Performance Bottlenecks

• O(n²) or worse algorithms
• Memory leaks
• Inefficient queries
• Scalability constraints

3. Report Generation

Each prediction includes:

• Specific code location: filename, line number, function name
• Issue description: why it will become a problem in the future
• Impact estimate: timeline and scope of impact
• Fix suggestion: prioritized preventive measures

Quick Start

Basic Usage Flow

1. Run analysis: Specify target and start analysis
2. Risk assessment: Review detected issues and risk levels
3. Decide response plan: Plan response based on priority
4. Track: Manage as Todo/Issue (optional)

Tracking Options After Analysis

After analysis completes, you can choose from the following tracking methods:

"How would you like to track these predictions?"

1. Create Todos: Track resolution progress
2. Create GitHub Issues: Generate issues with detailed information
3. Summary only: Provide actionable report without creating tasks

Output Example

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🔮 Predictive Code Analysis Report
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

[Critical] Security Risk
📍 Location: src/auth/login.ts:45-67
⚠️ Issue: User input concatenated into SQL string without input validation
📅 Timeline: Can be exploited immediately
💥 Impact: Risk of SQL injection attacks
🛠️ Mitigation: Use prepared statements or ORM

[High] Performance Bottleneck
📍 Location: src/api/search.ts:120-145
⚠️ Issue: O(n²) nested loop for data filtering
📅 Timeline: Latency occurs when data exceeds 1,000 records
💥 Impact: Response time over 10 seconds, degraded UX
🛠️ Mitigation: Switch to O(n) algorithm using Map/Set

Detailed References

For more detailed information, see:

• Risk Assessment Framework: Evaluation criteria and severity determination details
• Analysis Methodology: Static analysis and pattern detection implementation
• Issue Categorization: Classification details for technical debt/security/performance
• Mitigation Strategies: Fix patterns by category

Practical Examples

Real-world usage examples:

• Risk Report Examples: Actual analysis reports and response examples
• Project-Specific Analysis: Analysis patterns by framework/language
• Preventive Measures: Concrete preventive measures and implementation examples

Dependencies

• project-detector: Automatic project type detection
• MCP Serena: Code structure analysis and dependency tracking

Important Notes

This skill does NOT:

• Add signatures or watermarks indicating AI generation
• Change repository settings or permissions
• Automatically fix code without permission

All suggestions are subject to human review and approval.