Score Utilization Guide
Understanding COI, ORI, and GSS metrics for comprehensive code stability analysis
Overview
MistSeeker is a tool that analyzes the mathematical stability of code. It evaluates the structural and execution stability of code through three dimensions: COI (Code Organization Index), ORI (Operation Reliability Index), and GSS (Geometric Stability Score).
Each score measures different risk factors, and when combined, they provide a comprehensive, multi-dimensional understanding of your codebase's stability, safety, and consistency.
COI (Code Orthogonal Integrity)
✓ Meaning
COI evaluates how well-designed your code is based on function length, complexity, nesting depth, and structural density. It measures the structural integrity and organization of your codebase.
✓ Use Cases
- •Refactoring prioritization — Identify which code needs attention first
- •Bottleneck detection — Find structural bottlenecks in large-scale projects
- •Stability assessment — Evaluate stability and extensibility when adding or modifying features
- •Code review focus — Automatically select areas that need intensive review
⚠ When COI is Low
- • Long and complex functions
- • Deep nesting structures
- • Multiple duplicate logic patterns
- • Low structural consistency
→ Increased maintenance difficulty and higher bug risk
ORI (Operational Robustness Index)
✓ Meaning
ORI quantifies factors that make execution unstable, such as external I/O, global state changes, environment dependencies, and randomness. It measures the safety and predictability of execution paths.
✓ Use Cases
- •Early risk detection — Identify execution risks that may occur in production environments
- •Side-effect refactoring — Remove side effects to improve code quality
- •Security review — Analyze CWE/CVE vulnerability causes
- •Test coverage — Strengthen test coverage for unstable code paths
- •Non-deterministic flow detection — Detect unpredictable code execution patterns
⚠ When ORI is Low
- • External calls mixed within logic
- • Global state pollution
- • Random/time-dependent behavior
- • Different behavior depending on environment
→ Higher risk of failures, unpredictable behavior, and security vulnerabilities
GSS (Geometric Stability Score)
✓ Meaning
GSS measures how stably code semantics are maintained even with small perturbations, using embedding-based analysis. It evaluates semantic stability through: Perturbation changes → Embedding movement → Curvature calculation → Semantic stability assessment.
✓ Use Cases
- •AI/LLM code validation — Verify the risk level of AI-generated code
- •Refactoring verification — Determine if semantic meaning is preserved after refactoring
- •Semantic drift detection — Detect when code meaning shifts over time
- •Stability mapping — Build geometric stability maps for large-scale projects
⚠ When GSS is Low
- • Code meaning changes significantly with small modifications
- • Code exists in geometrically distorted regions
- • Frequently appears in AI-generated code or rushed patches
→ Increased maintenance difficulty and semantic instability
Score Combination Analysis
| Combination | Characteristics | Interpretation |
|---|---|---|
| COI ↑ / ORI ↑ / GSS ↑ | Fully stable | Most ideal structure |
| COI ↓ / ORI ↑ / GSS ↑ | Structural issues | Refactoring needed |
| COI ↑ / ORI ↓ / GSS ↑ | Execution risks | Operational and side-effect risks |
| COI ↑ / ORI ↑ / GSS ↓ | Semantic instability | Possible AI-generated code risks |
| COI ↓ / ORI ↓ / GSS ↓ | All dimensions unstable | High-risk code (priority inspection) |
Final Conclusion
COI
"Structural Health"
Use for refactoring and module design
ORI
"Execution Flow Safety"
Use for operational stability, security, and test enhancement
GSS
"Semantic Stability"
Use for LLM code quality and future maintenance risk assessment
These three scores address different problem dimensions. By analyzing all three axes together, you can quantify the overall stability of your codebase and make informed decisions about code quality, refactoring priorities, and risk management.