No requests
recorded yet
💻
Disconnected
● Offline
🔧 Debug Log
[INFO] Debug panel ready. Click
Diagnostics or Test API to begin.
⚠ Failure Intelligence
| Time | Type | Evidence | Strategy | Outcome | Actions |
|---|---|---|---|---|---|
| Click Refresh to load failure data | |||||
Summary
Total Failures—
Total Retries—
Retry Success—
Declined—
Top Failure Reasons
No data yet
🤖 Agent Dashboard
Polling
Failure Detector
Active
Detections0
Last Check—
Puppeteer
Active
Corrections0
Last Fix—
Proxy Hotpatcher
Active
Patches0
🔧 Three-Layer Hotpatch Control
Memory
0
patches
Byte-Level
0
patches
Server
0
patches
Event Timeline
0 events
—
READY
Click Refresh to load agent events
📊 Performance & Observability
Total Requests
0
requests
Avg Latency
—
milliseconds
Avg Tokens/s
—
t/s
Total Tokens
0
tokens generated
Latency Timeline
—
Latency
(ms)
—
Tokens/sec Timeline
—
Tokens/s
—
Model Benchmark Comparison
| Model | Requests | Avg Latency | Avg t/s | Best t/s | Total Tokens |
|---|---|---|---|---|---|
| No benchmark data yet — send some messages first | |||||
Recent Request History
0 requests
🛡️ Security & Hardening
CSP Directives
0
active policies
Blocked Inputs
0
sanitized
Rate Limit
OK
0 / 30 per min
URL Validation
✓
backend verified
🔒 Content Security Policy Audit
Click Refresh to audit CSP
⏱ Client Rate
Limiter
30 req/min
0 used
30 remaining
🧹 Input Validation & Sanitization
DOMPurify
Checking...
—
Max Input Length
1,000,000,000 chars
ENFORCED
Adjust Limit
Script Tag Filter
Blocks <script>, onerror=, onload=, javascript:
ACTIVE
URL Allowlist
localhost, 127.0.0.1
ENFORCED
Security Event Log
0
—
AUDIT
Security panel initialized — Phase 4 active
🔄 Backend Switcher
Switch between configured inference backends (Ollama, llama.cpp, remote APIs).
The active backend handles all
/api/generate and /v1/chat/completions requests.
Active Backend:
—
Click Refresh to load
backends
🚦 LLM Router
Intelligent prompt routing across backends. The router selects the optimal backend based on prompt
analysis,
model capabilities, latency history, and user preferences.
Router Status
—
Last Decision
—
Backend Capabilities:
No data yet
Routing Heatmap:
No heatmap data
Pinned Routes:
No pins
🐝 Swarm Dashboard
Distributed inference across multiple nodes. Monitor swarm health, manage workers,
and view task distribution in real time.
Swarm Status
—
Active Nodes
0
Running Tasks
0
Nodes:
No nodes connected
Recent Events:
No events
📝 Multi-Response Engine
Generate multiple response variants (Standard, Guided, Creative, eXpert) from a single prompt.
Compare and select preferred responses for quality tuning.
Engine Status
—
Total Generated
0
Templates (S/G/C/X):
No templates loaded
Preference Stats:
No preference data
🔍 ASM Debugger
Low-level debugger integration. View breakpoints, registers, memory, stack frames,
and disassembly from the running Win32IDE debug session.
Debug Status
—
Breakpoints
0
Threads
0
Registers:
No register data
Stack:
No stack data
Recent Events:
No events
🚨 Safety Monitor
Runtime safety layer. Monitors inference outputs for safety violations,
enables rollback, and provides content checking before delivery.
Safety Status
—
Violations
0
Recent Violations:
✔ No violations recorded
📋 Policy Engine
Agentic policy engine. Manages inference policies, heuristic suggestions, and runtime governance rules.
Routes: /api/policies, /api/policies/suggestions, /api/policies/apply, /api/policies/reject,
/api/policies/export, /api/policies/import, /api/policies/heuristics, /api/policies/stats.
Active Policies
0
Pending Suggestions
0
Heuristics
0
Policies:
Click Refresh to load...
Suggestions:
—
🔩 Hotpatch Layers
Three-layer hotpatch system: Memory (VirtualProtect), Byte-Level (mmap GGUF), Server (request/response
transform).
Routes: /api/hotpatch/status, /api/hotpatch/apply, /api/hotpatch/revert, /api/hotpatch/toggle.
Memory Layer
—
Byte Layer
—
Server Layer
—
Click Refresh to load hotpatch status...
🧠 Chain-of-Thought Inspector
Inspects the reasoning chain of the last agentic decision. Shows step-by-step logic from the explain
endpoint.
Routes: /api/agents/explain, /api/agents/explain/stats.
Total Explanations
0
Avg Steps
—
Click Refresh to load last explanation chain...
🛠 Tool Executor
Execute backend tools: read_file, write_file, list_directory, execute_command, git_status.
Route: POST /api/tool { "tool": "...", "args": {...} }.
Output will appear here...
🏭 Engine Explorer
Maps all 150+ C++ engine source files to their subsystem categories and API endpoints.
Click a subsystem to see its files and try its endpoints.
Click Refresh to load engine map...
📈 Server Metrics
Live server metrics: uptime, request counts, memory usage, model status.
Routes: /metrics, /health, /api/status, /status.
Uptime
—
Requests
0
Memory
—
Health
—
Click Refresh to load metrics...
🤖 Subagent & Chain
Spawn subagents or chain multiple agents for complex tasks.
Routes: POST /api/subagent, POST /api/chain.
Output will appear here...