# Reasoning Pattern Configuration Schema
This directory contains JSON configuration files that define domain-specific reasoning patterns.
Each file defines patterns for a specific domain (e.g., database, API, security).
## File Structure
Each JSON configuration file must follow this structure:
```json
{
"version": "1.0.0", // Required: Semver version (e.g., "1.0.0")
"domain": "domain-name", // Required: Domain identifier (e.g., "database")
"description": "...", // Required: Human-readable description
"patterns": [...], // Required: Array of pattern definitions
"testCases": [...] // Optional: Array of test cases for validation
}
```
## Pattern Definition
Each pattern in the `patterns` array must include:
```json
{
"id": "unique-pattern-id", // Required: Unique identifier
"name": "Pattern Name", // Required: Human-readable name
"description": "...", // Required: What this pattern detects
"indicators": [...], // Required: Array of indicators (min 1)
"negativeIndicators": [...], // Optional: Indicators that reduce match score
"hypotheses": [...], // Required: Array of hypothesis templates (min 1)
"recommendations": [...], // Required: Array of recommendation templates (min 1)
"severity": "medium", // Required: "critical" | "high" | "medium" | "low"
"qualityThreshold": 0.5 // Optional: Minimum quality score (0-1, default 0.5)
}
```
## Indicator Definition
Indicators trigger pattern matching. Each indicator must include:
```json
{
"type": "exact", // Required: "exact" | "fuzzy" | "regex"
"value": "search phrase", // Required: Pattern value to match
"weight": 0.8, // Required: Scoring weight (0-1)
"keyTermCategory": "..." // Optional: "domainTerms" | "actionVerbs" | "nounPhrases" | "terms"
}
```
### Indicator Types
- **exact**: Case-insensitive exact phrase matching
- **fuzzy**: All keywords must be present (any order)
- **regex**: Regular expression pattern matching
## Hypothesis Template
Hypotheses provide testable explanations for problems:
```json
{
"id": "hypothesis-id", // Required: Unique identifier
"statement": "...", // Required: Template with {{placeholders}}
"investigationSteps": [...], // Required: Array of concrete steps (min 1)
"expectedFindings": [...], // Required: Array of expected findings (min 1)
"relatedHypotheses": [...], // Optional: IDs of related hypotheses
"estimatedTime": "15-30 min", // Required: Time estimate string
"likelihood": 0.7 // Required: Base likelihood (0-1)
}
```
### Template Placeholders
Use these placeholders in hypothesis statements and recommendation actions:
- `{{primarySubject}}`: Main subject extracted from the problem
- `{{domainTerms}}`: Domain-specific terms from the problem
- `{{actionVerbs}}`: Action verbs from the problem
## Recommendation Template
Recommendations provide actionable steps:
```json
{
"id": "recommendation-id", // Required: Unique identifier
"type": "diagnostic", // Required: "diagnostic" | "remedial"
"action": "...", // Required: Template with {{placeholders}}
"tools": [...], // Optional: Array of relevant tools/commands
"expectedOutcome": "...", // Required: Expected result description
"prerequisites": [...], // Optional: Array of prerequisite steps
"priority": 8, // Required: Priority level (1-10, higher = more important)
"documentationLinks": [...] // Optional: Array of documentation URLs
}
```
### Recommendation Types
- **diagnostic**: Investigation/analysis actions
- **remedial**: Fix/resolution actions
## Test Cases
Test cases validate pattern matching:
```json
{
"id": "test-case-id", // Required: Unique identifier
"input": "problem text", // Required: Problem description to test
"context": "...", // Optional: Additional context
"expectedDomain": "domain", // Required: Expected domain match
"expectedPatternIds": [...], // Required: Expected pattern IDs (min 1)
"minConfidence": 0.7 // Required: Minimum confidence score (0-1)
}
```
## Directory Organization
Organize pattern files by category:
```
config/reasoning-patterns/
├── README.md # This file
├── technical/
│ ├── database.json
│ ├── api.json
│ ├── security.json
│ └── performance.json
├── business/
│ ├── customer-engagement.json
│ └── market-analysis.json
├── operations/
│ ├── project-management.json
│ └── team-dynamics.json
└── analysis/
├── root-cause-analysis.json
└── risk-assessment.json
```
## Example Configuration
See `example.json` in this directory for a complete working example.
## Validation
Pattern configurations are validated on load. Invalid files are skipped with detailed error messages.
Check the server logs for validation errors and warnings.
