HTHermes Trading Research Agent
EN日本語
GitHub
Built for part-time investors and solo traders

Turn Hermes Agent into your disciplined US equity trading assistant.

This profile distribution packages the vision of Claude Trading Skills into beginner-friendly Hermes workflows: pre-market prep, after-close review, market regime checks, earnings triage, portfolio risk review, and trade journaling.

Start the 10-minute setup See the routines
Human decision gate No auto-trading Degraded mode without optional keys Cron-ready routines

A trading copilot for process, not prediction.

Solo traders do not need another black-box signal feed. They need a repeatable operating system that remembers the checklist, calls the right research skills, records the thesis, and forces the final decision back to the human.

☀

Start the day prepared

Run a pre-market routine that checks macro events, market regime, earnings movers, breadth, and risk posture before you open the broker screen.

✓

Keep your discipline visible

Every candidate is framed with setup, thesis, invalidation, risk considerations, data freshness, and a human decision gate.

↺

Review without friction

After the close, turn scattered observations into a structured review: what changed, what invalidated, what to journal, and what to watch tomorrow.

Carrying forward the Claude Trading Skills vision.

The upstream project provides domain-specific market research skills. This repository does not fork that knowledge silently. It orchestrates those canonical skills through Hermes profiles, bundles, cron jobs, and safety gates so a trader can use them as a cohesive assistant.

Canonical skills stay upstream

Skill names and workflows are resolved from tradermonty/claude-trading-skills wherever possible.

Hermes adds the operating layer

Profiles, slash commands, scheduled briefs, tool routing, provider choice, and gateway delivery become part of your personal assistant.

Safety is product design

The profile is a research/process assistant: no hidden jobs, no order placement, no profit promises, and no signal-service UX.

One command for each recurring trading routine.

You do not need to memorize individual skill names. Use the bundle that matches the moment in your trading day.

/pre-market-routine

Macro calendar, regime, earnings movers, breadth, watchlist candidates, and the day’s risk gate.

/after-close-review

What changed today, sector rotation, open-trade review, journal prompts, and tomorrow preparation.

/market-regime-daily

A fast regime check for breadth, participation, and exposure ceiling.

/swing-opportunity-daily

Disciplined swing candidates with setup, thesis, invalidation, and risk assumptions.

/trade-ticket

A schema-validated trade ticket with DRAFT / REVIEW_READY / APPROVED / REJECTED / EXPIRED states. Ticket only; no execution.

/trade-journal

Turns trade notes into structured journal entries with thesis, invalidation, and risk record.

Designed for the reality of part-time trading.

The useful assistant is not the loudest one. It is the one that reduces context-switching when you only have a limited window before work, during lunch, or after the close.

US equitiesCANSLIM / Minervini-style workflowsEarnings and PEAD triageMarket breadthPortfolio heatTrade journaling

What it helps you avoid

Skipping the macro calendar. Chasing a gap-up without an invalidation level. Forgetting whether your exposure ceiling changed. Letting a thesis blur into a hope. Mixing research with execution.

What it helps you build

A calmer routine: research first, thesis second, risk third, human decision last.

Beginner-friendly setup path.

The easiest path is to install Hermes Agent first, then ask Hermes to do the repository setup for you. The manual commands stay below for people who prefer to see every step.

1

Easiest: ask Hermes to install this profile

After hermes setup, open a normal Hermes chat and paste the prompt on the right. Hermes can clone both repositories, install the profile alias, and tell you exactly which optional keys to add.

# Install and configure Hermes first
curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash
hermes setup
hermes

# Then paste this into Hermes:
Please install the Hermes Trading Research Agent profile from
https://github.com/tradermonty/hermes-trading-research-agent-work-package
using the upstream skills repo
https://github.com/tradermonty/claude-trading-skills

Use the alias trading-research-assistant, keep brokerage access paper/read-only,
and show me how to run /pre-market-routine when installation is complete.
Manual fallback: if you want a fully explicit install, or if your Hermes session does not have terminal/file tools enabled yet, use the step-by-step commands below.
1

Install and configure Hermes Agent

Use the official installer, then run the setup wizard to choose your model/provider.

2

Clone both repositories

The upstream repository holds canonical trading skills; this repository packages them into Hermes routines.

3

Install the profile alias

The alias trading-research-assistant gives you an isolated Hermes profile for trading research.

4

Run a routine

Start with /pre-market-routine, then enable cron jobs only when you intentionally want scheduled briefs.

# 1. Install Hermes Agent
curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash
hermes setup

# 2. Clone the trading skills and this profile
git clone https://github.com/tradermonty/claude-trading-skills.git
git clone https://github.com/tradermonty/hermes-trading-research-agent-work-package.git
cd hermes-trading-research-agent-work-package

# 3. Link upstream skills and install the profile
export CLAUDE_TRADING_SKILLS_REPO="$(realpath ../claude-trading-skills)"
hermes profile install "$(pwd)" --name trading-research-assistant --alias -y

# 4. Configure model/provider and optional market-data keys
trading-research-assistant model
cp ~/.hermes/profiles/trading-research-assistant/.env.EXAMPLE \
   ~/.hermes/profiles/trading-research-assistant/.env

# 5. Use it
trading-research-assistant chat
# then type: /pre-market-routine
Optional keys: FMP and FINVIZ improve data-backed research. Alpaca keys are optional and default to paper/read-only behavior. Missing keys produce explicit degraded-mode notes instead of pretending the data exists.

Make the routine come to you.

Hermes cron can deliver scheduled research briefs after you explicitly create jobs. This repository ships presets for pre-market, after-close, weekly, and monthly reviews.

No hidden schedules: jobs are declared in data/schedule-presets.yaml and are only created when you run the cron script.

export HERMES_PROFILE_CMD=trading-research-assistant
export HERMES_CRON_DELIVER=local
bash cron/create_cron_jobs.sh
trading-research-assistant cron list

# Gateway must be running for scheduled delivery
trading-research-assistant gateway install
trading-research-assistant gateway start
兼業投資家・ソロトレーダーのためのHermesプロファイル

Hermes Agentを、あなた専用の米国株トレード・リサーチアシスタントに。

このリポジトリは、Claude Trading Skills の思想を引き継ぎ、Hermes Agent上で「寄り付き前チェック」「引け後レビュー」「相場環境確認」「決算銘柄トリアージ」「ポートフォリオリスク確認」「トレード記録」をすぐ呼び出せる形にまとめたプロファイル配布パッケージです。

朝

朝の限られた時間をチェックリスト化

マクロ予定、相場環境、決算銘柄、ブレッド、ウォッチリスト候補、リスクゲートをまとめて確認できます。

守

売買判断の前に規律を挟む

候補には仮説、否定条件、リスク、データ鮮度、出典、人間の最終判断ゲートを必ず含めます。

録

引け後レビューと記録を習慣化

その日の変化、保有銘柄の仮説、明日の準備、トレードジャーナルを構造化できます。

これは「シグナル配信」ではなく、あなたの投資プロセスを支える道具です。

兼業投資家に必要なのは、煽るような売買指示ではなく、限られた時間で「見るべきものを見て、記録すべきことを記録し、最後は自分で判断する」ための仕組みです。

/pre-market-routine

寄り付き前の相場環境・マクロ・決算・ブレッド・候補銘柄・リスクゲート。

/after-close-review

引け後の変化確認、セクターローテーション、保有銘柄レビュー、翌日の準備。

/swing-opportunity-daily

リスクゲートが許す時だけ、仮説・否定条件・リスク前提つきでスイング候補を整理。

/trade-ticket

候補をDRAFT / REVIEW_READY / APPROVED / REJECTED / EXPIREDの状態を持つチケットとして記録。発注はしません。

Claude Trading Skillsの知見を、Hermesの「日々使える相棒」に変換する。

上流リポジトリのスキルを正本として尊重しながら、このリポジトリはHermesのprofile、slash command、cron、gateway、safety gateを使って、個人投資家が毎日使いやすい形に束ねます。

個別スキル名を覚えなくてよい

「寄り付き前」「引け後」「決算」「ジャーナル」など、行動単位のコマンドから始められます。

データが無い時は無いと明記

APIキー不足や外部リポジトリ不足はdegraded modeとして明示し、事実を装いません。

最終判断は常に人間

自動売買、利益保証、隠しcron、発注代行は対象外。研究・記録・リスクレビューの補助に徹します。

導入手順

いちばん簡単なのは、まずHermes Agentをインストールしてから、Hermes本人にこのプロファイルの導入を依頼する方法です。すべてのコマンドを自分で確認したい方向けに、従来の手動手順もそのまま下に残しています。

AI

最短: Hermesに導入を依頼する

hermes setup 後に通常のHermesチャットを開き、右の依頼文を貼り付けます。Hermesが2つのリポジトリをcloneし、profile aliasをinstallし、最後に使い方を案内します。

# まずHermes Agentをインストール・初期設定
curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash
hermes setup
hermes

# 次に、Hermesへこの依頼文を貼り付け:
以下のHermes Trading Research Agent profileを導入してください。
https://github.com/tradermonty/hermes-trading-research-agent-work-package

上流スキルリポジトリはこちらです。
https://github.com/tradermonty/claude-trading-skills

aliasは trading-research-assistant にしてください。
ブローカー連携はpaper/read-only前提のままにしてください。
導入後、/pre-market-routine の実行方法も教えてください。
手動手順: Hermesのterminal/file toolsがまだ使えない場合や、全ステップを自分で確認したい場合は、以下の従来手順を使ってください。
# 1. Hermes Agentをインストール
curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash
hermes setup

# 2. 上流スキルとこのプロファイルをclone
git clone https://github.com/tradermonty/claude-trading-skills.git
git clone https://github.com/tradermonty/hermes-trading-research-agent-work-package.git
cd hermes-trading-research-agent-work-package

# 3. 上流スキルの場所を指定してprofile install
export CLAUDE_TRADING_SKILLS_REPO="$(realpath ../claude-trading-skills)"
hermes profile install "$(pwd)" --name trading-research-assistant --alias -y

# 4. モデル/providerと.envを設定
trading-research-assistant model
cp ~/.hermes/profiles/trading-research-assistant/.env.EXAMPLE \
   ~/.hermes/profiles/trading-research-assistant/.env

# 5. 起動して寄り付き前ルーチンを実行
trading-research-assistant chat
# チャット内で: /pre-market-routine
GitHubで見る上流のClaude Trading Skillsを見る