Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

QStrader - AI Trading Assistant

v1.0.0

AI Trading Assistant for quantumstocks.ru. Automated hedge fund with market analysis, risk management, and trade execution via n8n MCP. Use when analyzing ma...

0· 108·0 current·0 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for antonby77/qstrader.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "QStrader - AI Trading Assistant" (antonby77/qstrader) from ClawHub.
Skill page: https://clawhub.ai/antonby77/qstrader
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install qstrader

ClawHub CLI

Package manager switcher

npx clawhub@latest install qstrader
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
medium confidence
!
Purpose & Capability
The skill claims to be a trading assistant using n8n MCP, which matches the code (mcporter calls, risk checks, logging). However the registry metadata declares no required environment variables or binaries even though SKILL.md and scripts expect .env entries (N8N_MCP_URL, QDRANT_URL, QDRANT_API_KEY, LIGHTRAG_USERNAME/PASSWORD, OPENROUTER_API_KEY) and require 'mcporter' and Python. This mismatch between declared requirements and actual needs is an incoherence.
!
Instruction Scope
SKILL.md and scripts instruct the agent to call many MCP endpoints including write endpoints (Place_Order, Close_an_open_deal). The documentation insists on human confirmation for writes, but nothing in the code enforces interactive confirmation — the agent or an automated workflow could call Place_Order if prompted. Scripts also read local mcporter config files (~/.openclaw/workspace/config/mcporter.json and similar) and reference instrument.json paths, which gives access to existing configured broker connections and their credentials if present.
Install Mechanism
There is no formal install spec (instruction-only), which lowers install-time risk. setup.sh is provided and suggests installing mcporter via npm; the skill will instruct users to run that script. No remote downloads or obfuscated installers are present in the package, but running setup.sh or the suggested npm install will modify the host environment and may register MCP endpoints.
!
Credentials
The package expects multiple secrets and endpoints (.env.example lists N8N_MCP_URL, QDRANT_URL/KEY, LightRAG creds, OPENROUTER_API_KEY) yet the skill metadata declares none. Requesting broker/control endpoints and database/LLM keys is reasonable for this purpose, but those are sensitive credentials and their absence from the declared requirements is misleading and reduces the ability to evaluate privilege scope before install.
Persistence & Privilege
always:false and normal model invocation behavior — no forced permanent inclusion. setup.sh may add a mcporter config entry (mcporter add my-n8n-mcp) into the user's mcporter configuration which is expected. The combination of autonomous invocation (platform default) plus the skill's ability to call write endpoints is worth attention, but autonomous invocation alone is not a disqualifier.
What to consider before installing
This skill appears to implement what it claims (market analysis, risk checks, and mcporter-based trading), but there are important mismatches and operational risks you should address before installing: 1) Metadata mismatch: The skill DOES expect and instruct you to populate a .env with N8N_MCP_URL, QDRANT and LightRAG credentials, and an OpenRouter key, yet the registry lists no required env vars. Don’t assume the registry has enumerated needed secrets — inspect and populate .env yourself. 2) Broker write ability: The skill can call Place_Order and other write endpoints via mcporter. SKILL.md asks for human confirmation, but the code does not technically enforce an interactive confirmation step. If you allow autonomous agent actions, an agent could place live orders. Limit agent autonomy or require an explicit, auditable manual confirmation step before any write call. 3) mcporter config and scope: The scripts auto-detect and use existing mcporter config files (~/.openclaw/workspace/config/mcporter.json, ~/.mcporter/mcporter.json). Before running, inspect those files to ensure they reference only broker/MCP endpoints you control; otherwise the skill could use pre-existing credentials unintentionally. 4) Run in a safe environment: Test thoroughly in a sandbox or paper-trading account (n8n MCP test/demo) before connecting to live brokers. Verify risk_manager behavior (it returns non-zero on rejection) and that your operational process prevents accidental live orders. 5) Review .env.example and sources: Fill only the minimal credentials required, rotate keys used for testing, and avoid putting high-privilege production keys into the skill until tested. Consider restricting OPENROUTER/QDRANT keys to limited-scope/test projects. 6) Operational checks: If you intend to use this skill, update the registry metadata to list required env vars and required binaries (mcporter, python3). Add or demand a programmatic confirmation step (or disable autonomous invocation) before write endpoints are called. If you want, I can produce a checklist you can run before enabling the skill or suggest a minimal .env and a safe confirmation wrapper to enforce manual approval.

Like a lobster shell, security has layers — review code before you run it.

latestvk971as785cbydk1ew6fs0qdx8h83ps66
108downloads
0stars
1versions
Updated 1mo ago
v1.0.0
MIT-0

QStrader — Торговый ассистент AI хедж-фонда

Архитектура

QStrader работает через n8n MCP — единый торговый терминал с 45+ инструментами. Доступ к MCP осуществляется через mcporter CLI или напрямую через MCP.

Пользователь → Агент → mcporter call → n8n MCP → Брокер / Данные

⚠️ Правила безопасности (критично!)

  1. SL/TP обязательно — ни одной позиции без стоп-лосса и тейк-профита
  2. Margin < 50% — если больше, немедленно закрываем худшие позиции
  3. Дневной убыток < 2% — превысил → стоп торговли на день
  4. Числа проверяй дважды — ошибка 640 вместо 6400 стоит дорого
  5. Защита > атака — закрытие убыточных позиций приоритетнее открытия новых
  6. Только с подтверждения — любой ордер требует согласия пользователя

Подробнее: references/risk-rules.md

Workflow: Анализ перед входом в сделку

Шаг 1. Получить данные аккаунта

mcporter call my-n8n-mcp.Get_account_data

Проверить: баланс, маржа, свободные средства.

Шаг 2. Технический анализ

mcporter call my-n8n-mcp.get_technical_analysis ticker=^GSPC

Как читать индикаторы:

  • RSI (0-100): <30 — перепроданность (pot buy), >70 — перекупленность (pot sell)
  • MACD: MACD > Signal → бычий, MACD < Signal → медвежий. Гистограмма показывает momentum
  • Bollinger Bands: цена у верхней полосы → перекуплен, у нижней → перепродан
  • EMA 9/21: EMA9 > EMA21 → uptrend, EMA9 < EMA21 → downtrend

Шаг 3. LSTM прогноз

mcporter call my-n8n-mcp.predict_future_price_lstm ticker=^GSPC start_date=2025-04-27-00-00 end_date=2026-03-28-00-00 interval=1h future_steps=10 time_step=512
  • start_date: ~11 месяцев назад (YYYY-MM-DD-HH-MM)
  • end_date: завтра
  • future_steps: количество будущих баров для прогноза
  • time_step: длина обучающего окна

Обратить внимание на: VaR, рекомендованные стоп-лоссы, направление прогноза.

Шаг 4. Новости и сентимент

mcporter call my-n8n-mcp.CNBC_news input="S&P 500"
mcporter call my-n8n-mcp.BBG_market input="stocks"

Шаг 5. Проверка risk limits

Используй scripts/risk_manager.py или проверь вручную:

  • Margin usage < 50%
  • Дневной убыток < 2%
  • Размер позиции разумный
  • SL/TP установлены

Шаг 6. Решение и подтверждение

Сформулируй thesis:

  • Направление + причина (TA + LSTM + новости)
  • Вход, SL, TP, объём
  • R:R ratio (минимум 2:1)
  • Запроси подтверждение у пользователя

Workflow: Открытие позиции

  1. Проверь risk limits (шаг 5 выше)
  2. Определи брокерский тикер из instrument.json (см. references/ticker-formats.md)
  3. Отправь ордер:
mcporter call my-n8n-mcp.Place_Order ticker=US500 side=buy type=market volume=0.1 price=0 stop_loss=6400 take_profit=6800
  1. Залогируй сделку:
python3 scripts/trade_logger.py US500 buy 6600 0.1 "EMA crossover + LSTM bullish" --tags "indices,trend"

Workflow: Закрытие позиции

mcporter call my-n8n-mcp.Close_an_open_deal deal_id=12345

Или закрыть конкретный тикер — сначала получить список сделок:

mcporter call my-n8n-mcp.Deals

Ticker форматы

КонтекстФорматПримеры
Брокерские ордераИз instrument.jsonUS500, XAUUSD, EURUSD, TSLA, VXX
Аналитика/новостиYahoo Finance^GSPC, ^DJI, ^IXIC, GC=F, EURUSD=X, TSLA
LSTM датыYYYY-MM-DD-HH-MM2025-04-27-00-00

Подробнее: references/ticker-formats.md

Ключевые MCP эндпоинты

Аккаунт и торговля

ИнструментНазначение
Get_account_dataБаланс, эквити, маржа
DealsТекущие открытые сделки
Place_OrderОткрыть позицию (⚠️ с подтверждения!)
Close_an_open_dealЗакрыть сделку по ID

Аналитика

ИнструментНазначение
get_technical_analysisRSI, MACD, Bollinger, EMA
predict_future_price_lstmLSTM прогноз + VaR
get_options_dataWalls, gamma, PCR, max pain
get_company_fundamentalsФундаментал компании
get_company_financial_scoresСкоринг 0-100

Новости

ИнструментНазначение
CNBC_newsНовости CNBC
BBG_marketBloomberg рынки
BEREZIN_sentimentСентимент Березина

Полный справочник: references/mcp-endpoints.md

Скрипты

СкриптНазначение
scripts/setup.shПервичная настройка (mcporter + .env)
scripts/market_analysis.pyЕдиный анализ актива (TA + LSTM)
scripts/risk_manager.pyПроверка risk limits перед ордером
scripts/trade_logger.pyЛог сделок в Qdrant

Первичная настройка

cd skills/qstrader
cp .env.example .env  # Заполнить свои ключи
bash scripts/setup.sh

Comments

Loading comments...