Complete guide to using the J4C Enhanced Agent Framework Portal
J4C (Jeeves for Coders) Enhanced Agent Framework is your comprehensive AI agent enhancement platform. This portal provides a visual interface to explore mental models, view analytics, test APIs, and manage agent integrations.
Navigate to http://j4c.aurigraph.io in your web browser.
The dashboard loads automatically, showing real-time statistics and system health.
Use the sidebar navigation to explore different sections of the portal.
The dashboard provides a real-time overview of the J4C framework status and usage.
Total number of available cognitive models (22 in v3.1)
Number of model categories (7 categories)
Total API requests processed today
Server uptime duration
Shows request volume over time. Use the dropdown to filter by time period (24 hours, 7 days, 30 days).
Pie chart showing usage distribution across model categories.
Displays current system health metrics:
Shortcut buttons to navigate to common tasks:
The Mental Models page displays all 22 cognitive frameworks organized by category.
Use the filter tabs at the top to show models by category:
Each model is displayed as a card showing:
Click on any model card to view detailed information including:
Use the search bar in the header to quickly find specific models by name or description.
This page explains the 7 reasoning strategies available in J4C for structured problem-solving.
| Strategy | When to Use |
|---|---|
| Forward Chain | When you have clear facts and need to derive conclusions |
| Backward Chain | When you know the goal but need to find the path |
| Bidirectional | Complex problems where both start and end are known |
| Depth First | When exploring one solution path thoroughly is important |
| Breadth First | When comparing multiple options at each level |
| Best First | When confidence scores can guide the search |
| Monte Carlo | For probabilistic problems with many possible paths |
Each reasoning chain consists of steps. Understanding these types helps interpret agent reasoning:
The Analytics page provides detailed usage statistics and performance metrics.
Shows request volume by hour, helping identify usage patterns and peak times.
Lists the most frequently called API endpoints with:
Use analytics to identify which mental models are most effective for your team's problem types.
Generate and download detailed reports in JSON format:
Contains API usage statistics, request trends, and endpoint popularity.
Mental model usage distribution, effectiveness scores, and category breakdown.
System health metrics, uptime statistics, and performance data.
Click on the report card you want to generate.
The report content will display in the panel below.
Click the "Download" button to save as JSON file.
Reports contain real-time data. Generate new reports to get the latest information.
The API Explorer allows you to test J4C API endpoints directly from the portal.
| Endpoint | Description |
|---|---|
/health |
Health check - verify server is running |
/api/v3/info |
Framework version and capabilities |
/api/v3/models |
List all mental models |
/api/v3/analytics |
Usage analytics data |
/api/v3/metrics |
Prometheus-format metrics |
/api/v3/docs |
Documentation index |
Click on any endpoint card to send a test request.
The response appears below showing status, time, and data.
The GNN (Graph Neural Network) Recommendations page provides AI-powered model suggestions based on your problem domain and context.
Choose from: Debugging, Architecture, Performance, Security, Refactoring, Planning, Team Organization, Testing, or Risk Assessment.
Narrow down with: Bug, Performance Issue, New Feature, Security Issue, Refactoring, Design Problem, Integration, or Optimization.
Enter keywords like "memory leak", "N+1 query", or "authentication" to improve recommendations.
Click "Get Recommendations" to receive ranked model suggestions with synergy information.
Each recommendation includes:
Click "Get Combinations" to find optimal combinations of 2-3 models for complex problems.
Pre-built chains for common scenarios:
The GNN engine learns from usage patterns. The more you use it, the better the recommendations become!
SPARC (Specification, Pseudocode, Architecture, Refinement, Completion) is a structured methodology for managing projects through 5 phases.
| Phase | Focus | Deliverables |
|---|---|---|
| Specification | Define what to build | Requirements, user stories, acceptance criteria |
| Pseudocode | Design the logic | Algorithms, flow diagrams, logic mapping |
| Architecture | Structure the system | Components, APIs, data models |
| Refinement | Build and iterate | Code, tests, bug fixes |
| Completion | Finalize and deploy | Documentation, deployment, monitoring |
Switch between views using the view selector:
Use SPARC phases to ensure nothing is missed. Don't rush through Specification and Pseudocode!
Agile sprint management for tracking work in time-boxed iterations.
The current sprint card shows:
Default view showing tasks in 4 columns:
Chart showing remaining work vs. ideal burndown line. Useful for tracking if the sprint is on track.
Bar chart comparing planned vs. completed story points across sprints.
List of all product backlog items with priority indicators:
| Points | Effort |
|---|---|
| 1 | Trivial (< 1 hour) |
| 2 | Small (1-2 hours) |
| 3 | Medium (half day) |
| 5 | Large (1 day) |
| 8 | Extra Large (2-3 days) |
| 13 | Epic - consider splitting |
If a task is 13 points or more, break it down into smaller, more manageable pieces.
AI agents can integrate with J4C to submit analysis data, receive recommendations, and learn from team experiences.
Contact your administrator to obtain an API key for agent authentication.
# API Key format
j4c_XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXPOST to /api/v3/agent/session to register the agent session.
GET /api/v3/agent/recommend for model suggestions based on domain.
GET /api/v3/agent/similar to find relevant past solutions.
POST to /api/v3/agent/analysis with your analysis results.
POST to /api/v3/agent/learning to share successful patterns.
PUT to /api/v3/agent/session/:id to close the session.
curl -X POST http://j4c.aurigraph.io/api/v3/agent/analysis \
-H "X-API-Key: j4c_YOUR_KEY" \
-H "Content-Type: application/json" \
-d '{
"agentId": "my-agent",
"problemDescription": "Memory leak in service",
"domain": "debugging",
"modelsUsed": ["first_principles", "systems_thinking"],
"insights": ["Object references not being released"],
"confidence": 0.85,
"outcome": "success"
}'See the full API Reference for detailed endpoint documentation.
| Error | Solution |
|---|---|
| 401 Unauthorized | API key required for protected endpoints |
| 403 Forbidden | Check permissions for your role |
| 404 Not Found | Endpoint doesn't exist - check URL |
| 500 Server Error | Server issue - try again later |
/health endpoint