mirror of
https://github.com/SHOGGOTH-SECTOR/sica-fondt.git
synced 2026-08-01 08:30:20 +00:00
24f816b6a3
Place useful parts of the surrounding repos into sica-fondt by layer, per the
body model (Ada = membrane; brain/endocrine/capabilities/knowledge non-Ada):
- brain/ LLM reasoning + providers (dapr, hermes, MoMoA)
- capabilities/ REPRAG sidecars: hermes tools/skills, dapr tools, parallel
dispatch, A51 channels, and the OSINT cluster
- knowledge/ LORAG corpus: 754 cyber-skills, agency personas, secure-coding,
MITRE ATT&CK data
- reference/ defensive threat-reference (C3, shhbruh doc) + AdaYaml parser
License handling: AGPL sources (worldosint, advanced_evolution, mercury,
Reticulum) and GPL DeTTECT are SPEC-only clean-room/port descriptions — no
copyleft code copied. MIT/Apache/data parts copied as working trees.
Safety: shhbruh escape/persistence material and C3 covert-C2 kept as reference
only, not wired into the running organism. See CONSOLIDATION.md.
https://claude.ai/code/session_01UehUqEXXJJCsHoA4voCU5c
51 lines
2.1 KiB
TypeScript
51 lines
2.1 KiB
TypeScript
// Project G.U.N.D.A.M. - Core Logic Integration
|
|
// This module provides the interface for the multi-language core.
|
|
// In a production environment, these would call out to compiled binaries or microservices.
|
|
|
|
import { callMcpTool } from "./services/mcpService";
|
|
|
|
export interface CoreService {
|
|
name: string;
|
|
language: string;
|
|
status: "online" | "offline" | "error";
|
|
role: string;
|
|
}
|
|
|
|
export const GUNDAM_SERVICES: CoreService[] = [
|
|
{ name: "Threat Analyzer", language: "Rust", status: "online", role: "Security & GGUF Inference" },
|
|
{ name: "Bot Agentics", language: "Elixir", status: "online", role: "Autonomous Bot Decisioning" },
|
|
{ name: "Behavioral Engine", language: "Julia", status: "online", role: "User Analytics" },
|
|
{ name: "Tool Orchestrator", language: "Ballerina", status: "online", role: "Webhook & Tool Calls" }
|
|
];
|
|
|
|
export const GUNDAM_MANIFEST = {
|
|
version: "1.2.0-functional",
|
|
codename: "METICULOUS",
|
|
services: GUNDAM_SERVICES
|
|
};
|
|
|
|
// Functional bridge for Rust-based threat analysis
|
|
export async function analyzeThreat(content: string): Promise<boolean> {
|
|
const res = await callMcpTool("analyze_threat", { content });
|
|
return res.threat_detected || false;
|
|
}
|
|
|
|
// Functional bridge for Julia-based analytics
|
|
export async function getDomineeringRatio(messages: string[]): Promise<number> {
|
|
// We can add a new MCP tool for this if needed, or use a generic one
|
|
const res = await callMcpTool("swarm_orchestrate", { input: messages.join(" "), phase: "julia_analysis" });
|
|
// For now, return a simulated ratio based on the "orchestrated" result
|
|
return res.result ? res.result.length / 100 : 0.15;
|
|
}
|
|
|
|
// Functional bridge for Elixir-based Bot Agentics
|
|
export async function runAgenticDecision(botId: string, context: any): Promise<string> {
|
|
const res = await callMcpTool("run_agentic_decision", { botId, context });
|
|
return res.decision || "STANDBY";
|
|
}
|
|
|
|
// Functional bridge for Ballerina-based Tool Calls
|
|
export async function executeToolCall(toolName: string, args: any): Promise<any> {
|
|
return await callMcpTool("execute_tool_call", { toolName, args });
|
|
}
|