feat: wire WEVIA+WEVCODE+OpsCenter skills (42 total)
This commit is contained in:
Yanis Mahboub
85
skills/public/weval-opscenter-manager/SKILL.md
Normal file
85
skills/public/weval-opscenter-manager/SKILL.md
Normal file
@@ -0,0 +1,85 @@
|
||||
---
|
||||
name: weval-opscenter-manager
|
||||
description: WEVAL OpsCenter Manager and infrastructure operations. 29 CAPS, 28 CLI tools, Sentinel relay chain, WEDROID backend agent, server management (S204/S95/S151), Docker, PMTA, PostgreSQL. Use when asked about WEVAL operations, server management, Sentinel, WEDROID, or infrastructure administration.
|
||||
---
|
||||
|
||||
# WEVAL OpsCenter Manager — Infrastructure Operations
|
||||
|
||||
## Relay Chain (Command Execution)
|
||||
```bash
|
||||
# CX relay (Claude → S204)
|
||||
CMD='your_command'
|
||||
CMD_B64=$(echo -n "$CMD" | base64 -w0)
|
||||
curl -sk -X POST "https://weval-consulting.com/api/cx" -H "Content-Type: application/x-www-form-urlencoded" -d "k=WEVADS2026&c=$CMD_B64"
|
||||
|
||||
# Droid relay (Claude → S204 → S95)
|
||||
curl -sk -X POST "https://weval-consulting.com/api/droid" -H "Content-Type: application/x-www-form-urlencoded" -d "k=DROID2026&c=$CMD_B64"
|
||||
|
||||
# S95 Sentinel direct
|
||||
curl http://10.1.0.3:5890/api/sentinel-brain.php?action=exec&cmd=CMD
|
||||
|
||||
# S95 Sentinel backup (when 5890 down)
|
||||
curl -sk -u weval:W3valAdmin2026 https://10.1.0.3:8443/api/sentinel-brain.php
|
||||
|
||||
# S151 SSH
|
||||
sshpass -p MX8D3zSAty7k3243242 ssh ubuntu@151.80.235.110
|
||||
```
|
||||
|
||||
## Servers
|
||||
| Server | IP | Role | SSH |
|
||||
|--------|-----|------|-----|
|
||||
| S204 | weval-consulting.com | Main (nginx, PHP, PG, Docker, Ollama) | root/FLVhtCX47w3e |
|
||||
| S95 | 95.216.167.89 | Backend (Apache, PMTA, Postfix, WEVADS) | root/rspmk4J9i3fm -p49222 |
|
||||
| S151 | 151.80.235.110 | Tracking (OVH, OpenClaw, relay) | ubuntu/MX8D3zSAty7k3243242 |
|
||||
|
||||
## WEDROID Agent (v5.0)
|
||||
- API: /api/wedroid-brain-api.php
|
||||
- Capabilities: diagnostic, correction, DB queries, cron management, service restart, deployment, scan, diff, bulk ops, nonreg tests
|
||||
- Chain: Claude → CX → S204 → S95 (droid)
|
||||
- RULE: ALWAYS WEDROID first for ALL backend operations
|
||||
|
||||
## Key Operations
|
||||
```bash
|
||||
# PostgreSQL S95
|
||||
ssh -p 49222 root@95.216.167.89 "PGPASSWORD=admin123 psql -U admin -h 127.0.0.1 -d adx_system -c 'SQL'"
|
||||
|
||||
# Docker S204
|
||||
sudo docker ps --format "table {{.Names}}\t{{.Status}}"
|
||||
|
||||
# PMTA
|
||||
ssh -p 49222 root@95.216.167.89 "pmta show status"
|
||||
|
||||
# Ollama models
|
||||
curl -s http://localhost:11434/api/tags
|
||||
|
||||
# DeerFlow services
|
||||
sudo systemctl status deerflow-langgraph deerflow-gateway deerflow-frontend
|
||||
|
||||
# Disk check (ALWAYS before deployment)
|
||||
df -h /
|
||||
|
||||
# Mattermost webhook
|
||||
curl -s -X POST http://localhost:8065/hooks/pt54hzthf3b6pe6rgp1ionipnh -H "Content-Type: application/json" -d '{"text":"message"}'
|
||||
```
|
||||
|
||||
## Docker Containers (S204)
|
||||
Authentik (server+worker+db), Plausible (+ClickHouse+PG16), Uptime-Kuma, Mattermost, SearXNG, Loki, Twenty-redis = 14 containers
|
||||
|
||||
## Docker Containers (S95)
|
||||
Vaultwarden:8222, n8n:5678, Qdrant:6333, Redis:6380
|
||||
|
||||
## Git Repositories
|
||||
- weval-consulting main (S204 /var/www/html)
|
||||
- weval-consulting i18n (S204 /var/www/weval)
|
||||
- wevia-brain master (S95)
|
||||
- wevads-platform main (S95 /opt/wevads)
|
||||
- deer-flow (S204 /opt/deer-flow)
|
||||
- GitHub: Yacineutt/weval-consulting.git (PAT auth)
|
||||
|
||||
## Critical Rules
|
||||
- GOLD backup before every modification, keep only latest, fix forward never rollback
|
||||
- NonReg PASS mandatory before every delivery
|
||||
- df -h BEFORE deployment (S204 hit 100% disk before)
|
||||
- Hex encoding for large file uploads (base64 fails)
|
||||
- NEVER sed -i on large PHP files → PHP str_replace + php -l
|
||||
- For complex shell escaping through relay: use php -r with chr() encoding
|
||||
42
skills/public/weval-wevcode-services/SKILL.md
Normal file
42
skills/public/weval-wevcode-services/SKILL.md
Normal file
@@ -0,0 +1,42 @@
|
||||
---
|
||||
name: weval-wevcode-services
|
||||
description: WEVCODE sovereign code assistant with 15 AI services. CodeAI, ProposalAI, BlueprintAI, SchemaAI, CyberGuard, AuditAI, EmailAI, MathSolver, VisionAI, and more. Use when asked about WEVCODE capabilities, service modes, or sovereign AI code assistant.
|
||||
---
|
||||
|
||||
# WEVAL WEVCODE — IA Code Assistant Souverain
|
||||
|
||||
## Access
|
||||
- URL: /wevcode or /wevia-fullscreen
|
||||
- File: /var/www/weval/wevia-ia/wevia-fullscreen.php (15KB)
|
||||
- Login required (SSO Authentik or credentials)
|
||||
- Model display: "WEVCODE IA - Souverain"
|
||||
|
||||
## 15 Service Modes
|
||||
| Mode | Service | Description |
|
||||
|------|---------|-------------|
|
||||
| chat | Chat libre | General AI conversation |
|
||||
| proposal | ProposalAI | Commercial proposals and devis |
|
||||
| blueprint | BlueprintAI | Architecture and technical blueprints |
|
||||
| schema | SchemaAI | Database schemas, ERDs, data models |
|
||||
| logo | LogoAI | SVG logo generation |
|
||||
| cyber | CyberGuard | Cybersecurity analysis and audit |
|
||||
| code | CodeAI | Code generation, review, debugging |
|
||||
| audit | AuditAI | IT audit and compliance review |
|
||||
| email | EmailAI | Professional email drafting |
|
||||
| translate | TranslateAI | Multi-language translation (FR/EN/AR/ZH) |
|
||||
| pdf | PDF Generator | Document generation from prompts |
|
||||
| vision | VisionAI | Image analysis and understanding |
|
||||
| math | MathSolver | Mathematical problem solving |
|
||||
| search | SearchAI | Web search via SearXNG |
|
||||
| planner | PlannerAI | Project planning and scheduling |
|
||||
|
||||
## Provider Stack
|
||||
- Primary: Groq llama-3.3-70b
|
||||
- Fallback: Cerebras → Mistral → SambaNova
|
||||
- Sovereign: Ollama S204 (granite4, qwen3:8b, meditron:7b)
|
||||
- Fast mode: granite4 on S151 (~8s response)
|
||||
|
||||
## Authentication
|
||||
- Login: admin-session-guard.php
|
||||
- SSO: Authentik (akadmin/YacineWeval2026)
|
||||
- Session: PHP sessions with auth gate
|
||||
72
skills/public/weval-wevia-public/SKILL.md
Normal file
72
skills/public/weval-wevia-public/SKILL.md
Normal file
@@ -0,0 +1,72 @@
|
||||
---
|
||||
name: weval-wevia-public
|
||||
description: WEVIA PUBLIC chatbot and Centre Commande capabilities. 71 modules, 20 AI providers, hybrid search (SearXNG+Qdrant+Ollama), vision, TTS, PDF generation, code execution, graph RAG. Use when asked about WEVAL chatbot, WEVIA features, provider chain, or Centre Commande modules.
|
||||
---
|
||||
|
||||
# WEVAL WEVIA PUBLIC — Chatbot + Centre Commande
|
||||
|
||||
## Architecture
|
||||
- Chatbot API: /var/www/weval/wevia-ia/weval-chatbot-api.php (305KB, ~5000 lines)
|
||||
- Widget: /var/www/html/wevia-widget.html (8.7KB)
|
||||
- Fullscreen: /var/www/weval/wevia-ia/wevia.html (Centre Commande)
|
||||
- Health: GET /api/weval-ia → {"status":"ok","engine":"WEVIA IA"}
|
||||
- Cache: /dev/shm/wevia_cache_* (TTL 5min)
|
||||
|
||||
## Provider Chain (20 providers)
|
||||
1. Groq llama-3.3-70b (default, fastest)
|
||||
2. Cerebras (Qwen-235B, fallback)
|
||||
3. Mistral (mistral-large)
|
||||
4. SambaNova
|
||||
5. Alibaba (qwen-plus)
|
||||
6-14. Ollama local S204 (9 models: glm4:9b, medllama2, meditron:7b, qwen3.5:0.8b, granite4, qwen3:8b, qwen3:4b, qwen2.5:7b, mistral)
|
||||
15-19. Ollama S151 (5 models)
|
||||
20. Manager v5 consensus (curl_multi parallel)
|
||||
|
||||
## Features
|
||||
- cot (chain of thought)
|
||||
- hybrid_search (SearXNG + Qdrant + KB)
|
||||
- smart_router (intent-based model routing)
|
||||
- learning_loop (RLHF feedback)
|
||||
- code_sandbox (code execution)
|
||||
- double_check (verified mode)
|
||||
- graph_rag (knowledge graph)
|
||||
- vision (image understanding)
|
||||
|
||||
## Modes
|
||||
- fast: sovereign (granite4 S151, ~8s)
|
||||
- deep: cloud providers with enrichment
|
||||
- verified: double-check with consensus
|
||||
- moa: mixture of agents (parallel query)
|
||||
|
||||
## Intent Routing (Sovereign Models)
|
||||
- code → granite4
|
||||
- schema → qwen3:8b
|
||||
- medical/ethica → meditron:7b
|
||||
- Chinese → glm4:9b
|
||||
- Arabic/Darija → qwen3:8b
|
||||
- translation → qwen3:8b
|
||||
|
||||
## Key Modules (71 total)
|
||||
Centre Commande includes: chat, PDF generation, vision, TTS, mermaid diagrams, code execution, cyber monitoring, LinkedIn content, knowledge base, graph RAG, calendar, multi-agent, memory, analytics dashboard, deploy, docker caps, failover monitor, email, translate, math solver, planner, image generation, OCR, browser automation, search, and more.
|
||||
|
||||
## APIs
|
||||
```bash
|
||||
# Health check
|
||||
curl -s https://weval-consulting.com/api/weval-ia
|
||||
|
||||
# Chat
|
||||
curl -s -X POST https://weval-consulting.com/api/weval-ia -H "Content-Type: application/json" -d '{"action":"chat","message":"Hello","mode":"deep"}'
|
||||
|
||||
# Capabilities
|
||||
curl -s -X POST https://weval-consulting.com/api/weval-ia -H "Content-Type: application/json" -d '{"action":"capabilities"}'
|
||||
|
||||
# Fast sovereign mode
|
||||
curl -s https://weval-consulting.com/fast.php -X POST -H "Content-Type: application/json" -d '{"message":"your question"}'
|
||||
```
|
||||
|
||||
## RULES
|
||||
- NEVER preg_replace [A-Z]\w+ on responses (eats everything) → use str_ireplace
|
||||
- NEVER sed via CX on PHP files → use Python str_replace
|
||||
- chattr -i BEFORE modifying protected files, +i AFTER
|
||||
- Reset opcache after API changes
|
||||
- Clear /dev/shm/wevia_cache_* after every API modification
|
||||
Reference in New Issue
Block a user