Execution Agent v3.1

Ship code.
Automate ops.
Monetize everything.

super 0pus is an autonomous AI agent that runs infrastructure, writes production code, manages crypto operations, and deploys from Telegram. No menus. No wizards. Just type what you want done.

super 0pus — Telegram
jerr@tg $ bikin landing page buat produk baru, deploy ke cf pages
⚡ Scanning skills...
→ interface-build-v3 (landing page)
→ cf-pages-ops (deploy)
→ anti-ai-slop-frontend (design QA)
📝 Writing HTML + Tailwind...
🎨 Applying dark theme, editorial layout...
🚀 wrangler pages deploy → done
✓ Live at your-project.pages.dev
✓ Custom domain app.jerrx.xyz (auto-TLS)
jerr@tg $ gas
18
Loadable Skills
7
Crypto Routes
<3s
Response Time
Session Memory

What it does

01
Full-Stack Code Generation
Next.js · React · Tailwind · Supabase · Docker
Writes production-ready code — not prototypes. Ships with imports, error handling, env examples, and deploy commands. Knows your stack, remembers your conventions, respects your anti-slop rules.
02
Infrastructure Automation
VPS · Nginx · PM2 · Systemd · Cloudflare
Bootstraps VPS, configures reverse proxies, sets up tunnels, deploys containers. Reads logs before blaming infra. Calculates latency math before recommending migrations.
03
Multi-Chain Crypto Ops
EVM · Solana · Sui · Aptos · TON
Swaps, bridges, DeFi positions, NFT minting, airdrop farming, mempool monitoring. Multi-provider RPC fallback. Nonce management. Simulate-before-broadcast. All from chat.
04
Browser Automation & RE
CloakBrowser · Playwright · HAR Capture · JsRpc
Captures network traffic, reverse-engineers CN-tier anti-bot systems, builds scrapers, injects fingerprints. 16 domain knowledge base with 151 techniques.
05
Artifact Generation
DOCX · XLSX · PPTX · PDF · Images
Generates documents, spreadsheets, presentations, and PDFs from templates or from scratch. ReportLab, python-docx, openpyxl — production output, not placeholder files.
06
Autonomous Self-Improvement
Skills · Memory · Fault Diagnosis · Self-Audit
Saves successful workflows as reusable skills. Persists memory across sessions. Diagnoses faults, applies fixes, and hardens against recurrence. Gets better every run.

Skill map — 18 loadable modules

m0
Skill Router
Routing meta, reflection loop, priority-weighted dispatch
m1
Monetization Ops
Pricing, cuan funnel, IDR + crypto monetization
m2
VPS Deploy
Ubuntu bootstrap, pm2, systemd, nginx config
m3
Signal Amplification
Content, viral mechanics, copy, campaign templates
m4
Telegram Bot Prod
Anti-duplicate, polling, webhook, multi-bot patterns
m5
Data Transform
Pandas pipelines, chunked reads, holder analysis
m6
API Bridge
Retry/timeout/idempotency, Midtrans, Xendit, WA
m7
LLM Orchestration
Multi-provider fallback, streaming, function calling
m8
Artifact Render
DOCX/XLSX/PPTX/PDF generation, image processing
m9
Interface Build
HTML+Tailwind, React+wagmi, conversion architecture
m10
Web3 On-Chain
RPC fallback, BIP39, NonceManager, mass farming
m11
Security Audit
Skill audit, Solidity flags, secret leak detection
m12
Batch Parallel
p-limit, asyncio sem, checkpoint, resume-from-fail
m13
Universal NFT Minter
Auto-detect mint fn, gas, OpenSea/Manifold/Zora
H1–H7
Hermes Crypto Agent
Swap, bridge, DeFi, sniping, monitoring, NFT, web3
x1
Self-Audit
Output QA, routing precision, token efficiency
x2
Strategic Decomposition
Reframe → decompose → options → pre-mortem → recommend
x3
Fault Diagnosis
Gather → classify → diagnose → resolve → verify → harden

How it works

"bikin bot telegram"
You type what you want
Natural language. Indonesian or English. No commands to memorize, no menus to navigate. Just say what you need.
auto-detect language
skill_router_v3()
Skills get loaded automatically
Keywords are scored against 18+ skills. The top match loads full context — API endpoints, proven workflows, pitfalls, deploy commands. Supporting skills load at 50%+ relevance.
priority-weighted match
execute_first()
Code ships before explanation
Working output comes first. Imports included. Error handling included. Run command included. Theory only if you ask. The reflection loop checks: "Is this immediately executable?"
execute first, explain after
memory.write()
It remembers. It improves.
Decisions get locked. Preferences persist. Successful workflows become skills. Failures get diagnosed and hardened. Every session makes the next one better.
persistent cross-session memory

Operating doctrine

"Is this answer immediately executable? If no → rewrite before sending."
super 0pus
Core reflection loop — every output, every time

Runs on

Hermes Agent
🧠
Claude Opus
🔷
Telegram
☁️
Cloudflare
🐳
Docker
🟢
Node.js
🐍
Python
⛓️
Multi-Chain
🦊
Solidity
🐧
Linux
📡
WebSocket
🔒
Self-Healing

Ready?

Stop explaining.
Start shipping.

super 0pus runs on Telegram. Open a chat, type what you need, and watch it execute.

Launch super 0pus →