Orchestration Console

I've analyzed the repository structure. Here's what I found:

The codebase follows a hexagonal architecture with clear separation between domain, infrastructure, and application layers. The agent system supports:

• YAML-based configuration with hot-reload
• MCP server integration for tool extensibility
• HITL (Human-in-the-loop) validation gates
• SSE streaming for real-time responses
• Subagent delegation for complex workflows

Would you like me to generate a detailed audit report?

Here is the Python utility I generated to parse the audit output. You can drop it into src/utils/audit_parser.py.

import json
from pathlib import Path
from typing import Any

def load_audit(path: Path) -> dict[str, Any]:
    with path.open("r", encoding="utf-8") as f:
        return json.load(f)

def summarize(audit: dict[str, Any]) -> dict[str, Any]:
    return {
        "agents": audit.get("agents", 0),
        "warnings": len(audit.get("warnings", [])),
        "critical": len([w for w in audit.get("warnings", []) if w.get("severity") == "critical"]),
    }

if __name__ == "__main__":
    result = summarize(load_audit(Path("audit.json")))
    print(f"Audit summary: {result}")

Use python -m utils.audit_parser to run it locally.