메뉴
Read any common document/data file — PDF, Word (.docx), Excel (.xlsx/.xls), PowerPoint (.pptx), images (OCR), CSV/TSV, plain text, JSON/YAML/TOML, HTML/XML, and most source-code files. Use the `read_document` tool.
기본적으로 소스를 먼저 확인하는 Prompt가 선택됩니다. 직접 명령으로 전환하거나 로컬 사본을 다운로드할 수도 있습니다.
설치 여부를 결정하기 전에 SKILL.md와 SkillsMP에 표시된 보조 파일을 읽어 보세요.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
직접 명령은 검토 Prompt를 거치지 않습니다. 실행하기 전에 소스를 확인하세요.
npx skills add https://github.com/HKUDS/Vibe-Trading --skill doc-reader명령은 한 줄로 유지됩니다. 복사하기 전에 가로로 스크롤해 전체 내용을 확인하세요.
로컬 사본을 원하시나요? SkillsMP에서 현재 제공할 수 있는 파일을 다운로드하세요.
SKILL.md 표시 중
Professional finance research toolkit — backtesting (9 engines + benchmark comparison panel), factor analysis, Alpha Zoo (462 pre-built alphas across qlib158/alpha101/gtja191/academic/fundamental), options pricing, 89 finance skills, 30 multi-agent swarm teams, Trade Journal analyzer, and Shadow Account (extract → backtest → render) across 24 market-data sources (tushare, yfinance, okx, binance, akshare, baostock, tencent, mootdx, ccxt, futu, mt5, local, eastmoney, sina, stooq, yahoo, pykrx, india_broker, qveris, longbridge, plus optional-key finnhub/alphavantage/tiingo/fmp).
The single ROUTER for every data need. Load this skill BEFORE any backtest, data-fetch, or research task to pick the best available source/tool, honour auth (env) requirements, and avoid ban-risk providers.
固收与信用分析：信用债评级、利差分析、违约风险评估、城投债研究、可转债定价与策略。
SOC 직업 분류 기준
| name | doc-reader |
| description | Read any common document/data file — PDF, Word (.docx), Excel (.xlsx/.xls), PowerPoint (.pptx), images (OCR), CSV/TSV, plain text, JSON/YAML/TOML, HTML/XML, and most source-code files. Use the `read_document` tool. |
| category | tool |
Return extracted text from any supported file in a single unified JSON envelope. The tool dispatches by file extension — you always call the same tool regardless of format.
| Category | Extensions | Notes |
|---|---|---|
.pdf | Text pages extracted in ms; scanned/image pages fall back to OCR | |
| Word | .docx | Paragraphs + table cells |
| Excel | .xlsx, .xls | All sheets, first 100 rows per sheet as preview |
| PowerPoint | .pptx | Slide text content |
| Images | .png/.jpg/.jpeg/.gif/.bmp/.webp/.tiff | OCR only |
| CSV / TSV | .csv, .tsv | Raw text with encoding fallback |
| Plain text | .txt/.md/.log/.rst | Encoding fallback |
| Config | .json/.yaml/.yml/.toml/.ini/.cfg/.env |
| Raw text |
| Markup | .html/.htm/.xml | Raw text (no HTML stripping) |
| Source code | .py/.js/.ts/.tsx/.go/.rs/.java/.cpp/.c/.sql/.sh/... | Raw text |
| Unknown extension | anything else | Best-effort read as UTF-8/GBK text |
Blocked (rejected at /upload): executables (.exe/.dll/.so/...) and
archives (.zip/.tar/...). Ask the user to unpack archives locally first.
Always call the tool directly — do not run Python from bash.
read_document(file_path="uploads/paper.pdf")
read_document(file_path="uploads/annual_report.pdf", pages="1-10")
read_document(file_path="uploads/contract.docx")
read_document(file_path="uploads/sales.xlsx")
read_document(file_path="uploads/deck.pptx")
read_document(file_path="uploads/chart.png") # image → OCR
read_document(file_path="uploads/config.yaml")
read_document(file_path="uploads/notes.md")
The pages parameter only applies to PDF; other formats ignore it.
All formats share this shape:
{
"status": "ok",
"file": "paper.pdf",
"format": "pdf",
"char_count": 52000,
"truncated": true,
"text": "..."
}
Format-specific extra fields:
| Format | Extra keys |
|---|---|
pdf | total_pages, pages_read, ocr_pages, ocr_engine, ocr_quality, skipped_pages |
docx | paragraphs, tables |
excel | sheets (array of {name, rows, cols}) |
pptx | slides |
text | encoding, size |
Content longer than 15000 chars is truncated; for PDFs use the pages
parameter to read slices.
1. read_document(file_path="paper.pdf") → full text
2. Extract abstract, methodology, conclusion → summarize
1. read_document(file_path="contract.docx") → paragraphs + tables
2. Flag key clauses (termination, liability, payment, IP)
1. read_document(file_path="sales.xlsx") → all sheet previews
2. If user wants trade journal analysis specifically, pivot to
`analyze_trade_journal` tool instead (see trade-journal skill).
1. read_document(file_path="scan.png") → OCR text
2. If OCR returns empty, tell the user; don't fabricate.
The read_document tool automatically uses OCR for PDF pages with insufficient extractable text.
Use min_text_per_page to control when OCR is triggered (default: 50 characters):
read_document("scanned_report.pdf", min_text_per_page=10) # More aggressive OCR
read_document("mixed_pdf.pdf", min_text_per_page=100) # Less aggressive OCR
Two OCR engines are built in — no extra packages needed beyond the engine SDK:
| Engine | Type | Requires | Install |
|---|---|---|---|
rapid | Local (offline) | rapidocr_onnxruntime | pip install rapidocr_onnxruntime |
llm-vision | Cloud | A vision-capable LLM model + API key | No extra install — uses your existing LLM provider config |
The llm-vision engine works with any OpenAI-compatible vision model (GPT-4o, Qwen-VL, Gemini, Claude, GLM-4V, etc.). It reuses your existing LANGCHAIN_PROVIDER / LANGCHAIN_MODEL_NAME / API key configuration — no separate provider mapping needed. If you explicitly set VIBE_TRADING_OCR_ENGINE=llm-vision, your model choice is trusted; a real API error from the provider is clearer feedback than a heuristic guess.
To override the model used for OCR (without changing your agent's main model):
VIBE_TRADING_OCR_LLM_MODEL=qwen3.7-plus
Set VIBE_TRADING_OCR_ENGINE to select the engine:
auto (default): use local engines only, never cloud (privacy: document pages never leave the machine)rapid: force RapidOCR (local, ONNX)llm-vision: force LLM vision OCR (cloud — pages are sent to your configured LLM provider)none: disable OCR entirelyPDF responses include OCR metadata:
ocr_engine: name of the OCR engine used (e.g. "rapid", "llm-vision") or nullocr_pages: number of pages processed via OCRskipped_pages: number of pages skipped (no OCR engine available)ocr_quality: object with quality_flag (good/degraded/no_ocr_engine/no_ocr_needed), ocr_pages, and text_density (chars per page)text
with a note field — tell the user to install rapidocr-onnxruntime or
set VIBE_TRADING_OCR_ENGINE=llm-vision with a vision-capable model.analyze_trade_journal instead.