Agent skill

Document Direct Python

by HKUDS in HKUDS/OpenSpace

Use direct Python execution for reliable document creation including spreadsheets, PDFs, and structured reports

MITAuto-check passedDocuments & Office

Install Document Direct Python

skills CLI
$ npx skills add HKUDS/OpenSpace --skill document-direct-python -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install HKUDS/OpenSpace document-direct-python --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/HKUDS/OpenSpace.git skills-src && mkdir -p .claude/skills && cp -r skills-src/benchmarks/gdpval/skills/spreadsheet-direct-python-merged .claude/skills/document-direct-python && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
document-direct-python
GitHub stars
7.7k
Token cost
~2.7k tokens
SKILL.md length
766 words
Files
2
Skills in repo
199
Repo updated
First seen
Licence
MIT

At a glance

Use direct Python execution for reliable document creation including spreadsheets, PDFs, and structured reports

  • Works in 4 steps: Assess what you need: Determine if the… → Proceed with available knowledge: For… → Document assumptions: Clearly note that… → …
  • Tasks that involve Excel spreadsheets
  • SKILL.md covers When to Use This Skill, Why Direct Execution?, How to Use and Spreadsheet Examples, plus 6 more sections
  • Calls python3

What it does

Document Direct Python is an agent skill from HKUDS/OpenSpace. Use direct Python execution for reliable document creation including spreadsheets, PDFs, and structured reports

Its SKILL.md is about 2.7k tokens, which your agent loads only when the skill is triggered. The skill folder holds 1 other file.

It sits in Documents & Office, covering Excel spreadsheets and PDF. It works with Python. The repository describes itself as: "OpenSpace: The Skill Management Layer for AI Agents" -- https://open-space.cloud/. The licence is MIT.

When your agent uses it

  • Tasks that involve Excel spreadsheets
  • Tasks that involve PDF

Example prompts

  • “/document-direct-python”

Requirements

  • Python 3

Workflow steps

4 steps, taken from the first numbered list in SKILL.md.

  1. Assess what you need: Determine if the information is from well-documented regulatory frameworks, standards, or common knowledge that you…
  2. Proceed with available knowledge: For pharmacy compliance, OSHA standards, GDPR requirements, and similar well-established frameworks…
  3. Document assumptions: Clearly note that materials were generated based on standard practices when source verification was unavailable.
  4. Use direct Python execution: Continue with the recommended run_shell pattern to generate documents even without web-sourced content.

What it can do on your machine

Read from SKILL.md and the folder at commit 3827781. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • python3

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Document Direct Python loads about 2.7k tokens when it runs. Until then it costs about 34 tokens; SKILL.md has 766 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~34
When it runs · the whole SKILL.md, loaded when a task matches
~2.7k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from HKUDS/OpenSpace at commit 3827781, republished under its MIT licence (© HKUDS). 766 words, ~2,722 tokens.

Download SKILL.mdSave it as .claude/skills/document-direct-python/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
document-direct-python
description
Use direct Python execution for reliable document creation including spreadsheets, PDFs, and structured reports

Direct Python Execution for Document Generation Tasks

When to Use This Skill

Use direct run_shell with Python scripts for document operations when:

  • Reading or writing complex Excel files with multiple sheets
  • Generating PDF documents, checklists, or reports
  • Applying formulas, formatting, or data transformations
  • Working with openpyxl, pandas, reportlab, FPDF, or similar libraries
  • The operation involves multiple steps that could exceed agent step limits
  • You need precise control over error handling and debugging
  • Complex scripts benefit from file-based execution for better reliability

Why Direct Execution?

The shell_agent tool can:

  • Hit maximum step limits on complex multi-step operations
  • Produce unexplained errors on formatting operations
  • Fail on intricate document reads/writes due to iterative parsing
  • Fail to parse heredoc syntax correctly, causing 'unknown error' failures

Direct run_shell with Python is more reliable because it:

  • Executes in a single step with no iteration limits
  • Provides clearer, immediate error messages
  • Handles complex operations without step constraints
  • Gives full control over library imports and execution flow
  • Writing scripts to .py files first avoids shell_agent parsing issues with heredocs

How to Use

Handling Web Research Failures

When web research tools (search_web, read_webpage) fail or return 'unknown error':

  1. Assess what you need: Determine if the information is from well-documented regulatory frameworks, standards, or common knowledge that you can provide from internal knowledge.
  2. Proceed with available knowledge: For pharmacy compliance, OSHA standards, GDPR requirements, and similar well-established frameworks, generate documents using your training knowledge rather than waiting for web access.
  3. Document assumptions: Clearly note that materials were generated based on standard practices when source verification was unavailable.
  4. Use direct Python execution: Continue with the recommended run_shell pattern to generate documents even without web-sourced content.

Example contingency workflow for pharmacy compliance:

python
# When web research fails, proceed with established regulatory knowledge
from reportlab.lib.pagesizes import letter
from reportlab.pdfgen import canvas

checklist_items = [
    "Verify pharmacist license is current and displayed",
    "Maintain controlled substance inventory logs",
    "Ensure proper storage temperatures for medications",
    "Keep patient counseling records for controlled substances",
    "Display required pharmacy signage and notices"
]
# Generate PDF using these standard compliance items

This approach was successfully used in task 045aba2e-4093-42aa-ab7f-159cc538278c_phase2 where all web tools failed but pharmacy compliance PDFs were still created successfully.

For complex multi-line scripts, especially when using shell_agent as executor:

bash
# Step 1: Write the Python script to a file
cat > generate_document.py << 'EOF'
import openpyxl
from openpyxl import Workbook

# Your document code here
wb = openpyxl.load_workbook('file.xlsx')
# ... operations ...
wb.save('output.xlsx')
print('Success')
EOF

# Step 2: Execute the script
python3 generate_document.py
Alternative Pattern: Inline Heredoc (Simple Scripts Only)

For short, simple scripts when NOT using shell_agent as the executor:

bash
python3 << 'EOF'
import pandas as pd
df = pd.read_excel('input.xlsx')
df.to_excel('output.xlsx', index=False)
print('Done')
EOF

Spreadsheet Examples

Example 1: Read and Transform Excel Data

Write to file first, then execute:

python
import pandas as pd

# Load data from specific sheet
df = pd.read_excel('input.xlsx', sheet_name='Revenue')

# Apply transformations
df['Net_Revenue'] = df['Gross_Revenue'] * (1 - df['Tax_Rate'])

# Save results
df.to_excel('output.xlsx', index=False, sheet_name='Processed')
Example 2: Multi-Sheet Operations with openpyxl
python
from openpyxl import load_workbook

wb = load_workbook('tour_data.xlsx')

# Iterate through sheets
for sheet_name in wb.sheetnames:
    ws = wb[sheet_name]
    # Apply formatting or calculations
    for row in ws.iter_rows(min_row=2, max_col=5):
        # Process cells
        pass

wb.save('tour_data_processed.xlsx')

PDF Generation Examples

Example 3: Generate PDF Checklist with reportlab

Write to file first, then execute:

python
from reportlab.lib.pagesizes import letter
from reportlab.pdfgen import canvas
from reportlab.lib.units import inch

def create_checklist(filename, items):
    c = canvas.Canvas(filename, pagesize=letter)
    width, height = letter
    
    # Title
    c.setFont("Helvetica-Bold", 20)
    c.drawString(1*inch, height - 1*inch, "Task Checklist")
    
    # Items
    c.setFont("Helvetica", 14)
    y_position = height - 1.5*inch
    for i, item in enumerate(items, 1):
        checkbox = "☐"  # Empty checkbox
        c.drawString(1*inch, y_position, f"{checkbox} {item}")
        y_position -= 0.3*inch
    
    c.save()
    print(f"Created {filename} with {len(items)} items")

# Usage
items = ["Review requirements", "Complete analysis", "Submit report", "Follow up"]
create_checklist("checklist.pdf", items)
Example 4: Generate PDF Report with FPDF
python
from fpdf import FPDF

class PDF(FPDF):
    def header(self):
        self.set_font('Arial', 'B', 15)
        self.cell(0, 10, 'Monthly Report', 0, 1, 'C')
        self.ln(10)
    
    def footer(self):
        self.set_y(-15)
        self.set_font('Arial', 'I', 8)
        self.cell(0, 10, f'Page {self.page_no()}', 0, 0, 'C')

def create_report(filename, data):
    pdf = PDF()
    pdf.add_page()
    pdf.set_font('Arial', '', 12)
    
    # Add content
    for section, content in data.items():
        pdf.set_font('Arial', 'B', 12)
        pdf.cell(0, 10, section, 0, 1)
        pdf.set_font('Arial', '', 12)
        pdf.multi_cell(0, 8, content)
        pdf.ln(5)
    
    pdf.output(filename)
    print(f"Report saved to {filename}")

# Usage
data = {
    "Executive Summary": "This report covers Q4 performance metrics.",
    "Key Findings": "Revenue increased by 15% compared to Q3.",
    "Recommendations": "Continue current strategy with minor adjustments."
}
create_report("report.pdf", data)
Example 5: Combined Spreadsheet to PDF Workflow
python
import pandas as pd
from reportlab.lib.pagesizes import letter
from reportlab.pdfgen import canvas
from reportlab.platypus import SimpleDocTemplate, Table, TableStyle
from reportlab.lib import colors

# Step 1: Process spreadsheet data
df = pd.read_excel('sales_data.xlsx')
summary = df.groupby('Region')['Revenue'].sum().reset_index()

# Step 2: Generate PDF report
doc = SimpleDocTemplate("sales_report.pdf", pagesize=letter)
elements = []

# Create table from data
data = [['Region', 'Revenue']]
for _, row in summary.iterrows():
    data.append([row['Region'], f"${row['Revenue']:,.2f}"])

table = Table(data)
table.setStyle(TableStyle([
    ('BACKGROUND', (0, 0), (-1, 0), colors.grey),
    ('TEXTCOLOR', (0, 0), (-1, 0), colors.whitesmoke),
    ('ALIGN', (0, 0), (-1, -1), 'CENTER'),
    ('GRID', (0, 0), (-1, -1), 1, colors.black),
]))

elements.append(table)
doc.build(elements)
print("PDF report generated successfully")

Error Handling Pattern

Write to file first, then execute:

python
import sys
from openpyxl import load_workbook

try:
    wb = load_workbook('data.xlsx')
    ws = wb.active
    
    # Your operations here
    value = ws['A1'].value
    
    wb.save('output.xlsx')
    print(f"Success: Processed {ws.max_row} rows")
    
except Exception as e:
    print(f"Error: {str(e)}", file=sys.stderr)
    sys.exit(1)

Best Practices

  1. Prefer file-based execution for complex scripts: write to .py file first, then execute via run_shell
  2. Import only needed libraries to reduce execution time
  3. Print clear success/error messages for debugging
  4. Save intermediate results for complex multi-step transformations
  5. Test with small data before scaling to large documents
  6. Use pandas for data manipulation and openpyxl for spreadsheet formatting
  7. Use reportlab for professional PDF layouts and FPDF for simple PDFs
  8. Clean up temporary script files after execution if they won't be reused
Show full SKILL.md (278 more words)Show less

When NOT to Use This Skill

  • Simple single-cell reads/writes (use shell_agent or basic commands)
  • Operations that require interactive user input
  • Tasks where you need the agent to iteratively refine the approach
  • Basic text file operations (use shell commands directly)

Common Libraries

Spreadsheets
LibraryBest For
openpyxlReading/writing .xlsx files, formatting, formulas
pandasData manipulation, analysis, merging datasets
xlrdReading older .xls files (read-only)
xlsxwriterCreating new .xlsx files with advanced formatting
PDF Generation
LibraryBest For
reportlabProfessional PDF reports with complex layouts
FPDFSimple PDF generation with basic formatting
PyPDF2 / pypdfReading, merging, splitting existing PDFs
pdfplumberExtracting text and tables from PDFs

Troubleshooting

Issue: Heredoc syntax fails with 'unknown error' when using shell_agent

  • Solution: Write the Python script to a .py file first, then execute it with python3 script.py. This pattern is significantly more reliable than inline heredoc execution when shell_agent is the executor.

Issue: FileNotFoundError

  • Solution: Verify the file path is absolute or relative to the working directory. Use os.getcwd() to check current directory if needed.

Issue: PermissionError

  • Solution: Ensure the file is not open in another application. Close Excel/PDF readers before writing.

Issue: MemoryError on large files

  • Solution: Process data in chunks using pandas chunksize parameter. For PDFs, generate in sections and merge.

Issue: Formatting not applying

  • Solution: Ensure you're modifying cell styles before saving, and use .copy() for style objects in openpyxl.

Issue: PDF text rendering issues

  • Solution: For non-ASCII characters, use appropriate fonts (e.g., DejaVuSans for Unicode support in reportlab).

Issue: Agent fails before first iteration (0 steps)

  • Solution: Ensure the skill is properly loaded and the task description clearly indicates document generation needs. Complex initialization may require explicit Python script patterns.

© HKUDS, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 1 other file in benchmarks/gdpval/skills/spreadsheet-direct-python-merged of HKUDS/OpenSpace.

  • SKILL.md
  • .skill_id

Open the folder on GitHubat commit 3827781

Compare with similar skills

Document Direct Python next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Document Direct Python compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Document Direct Python this skillHKUDS/OpenSpace7.7k—~2.7kAutomated safety check: PassMIT
Instrument Data To Allotropeaws-samples/amazon-bedrock-agents-healthcare-lifesciences2742 repos~2.7kAutomated safety check: PassApache-2.0
Doc Cleanernotoriouslab/doc-cleaner309—~712Automated safety check: PassMIT
MineruNebutra/MinerU-Skill122—~504Automated safety check: PassMIT
Office To Mdshuyu-labs/WebCode278—~1kAutomated safety check: NotesCustom licence
Exam IngestZeKaiNie/universal-examprep-skill301—~5.6kAutomated safety check: PassMIT

Similar skills

  • Instrument Data To Allotrope

    aws-samples/amazon-bedrock-agents-healthcare-lifesciences

    Official

    Convert laboratory instrument output files (PDF, CSV, Excel, TXT) to Allotrope Simple Model (ASM) JSON format or flattened 2D CSV.

    274 GitHub starsUsed in 2 repos~2.7k tokens
    Documents & OfficeAuto-check passed
  • Doc Cleaner

    notoriouslab/doc-cleaner

    Convert PDF, DOCX, XLSX, and text files to clean, structured Markdown.

    309 GitHub stars~712 tokensUpdated 1 mo ago
    Documents & OfficeAuto-check passed
  • Mineru

    Nebutra/MinerU-Skill

    An AI-Native skill for parsing PDF / Office / image files into Markdown with MinerU — a fast, zero-config document parser for AI agents.

    122 GitHub stars~504 tokensUpdated 13 days ago
    Documents & OfficeAuto-check passed
  • Office To Md

    shuyu-labs/WebCode

    Convert Office documents (Word, Excel, PowerPoint, PDF) to Markdown format.

    278 GitHub stars~1k tokensUpdated 3 mo ago
    Documents & OfficeAuto-check: notes
  • Exam Ingest

    ZeKaiNie/universal-examprep-skill

    从学生上传的课件/大纲/老师勾的重点/真题,一键初始化并验证备考工作区:解析 PDF、DOCX、PPTX、 XLSX、常见独立图片与 txt/md,建立分章节 LLM Wiki、标准题库、结构化接管队列与进度状态;仅在 Python 确实无法运行时 明确降级为手动写盘。当工作区尚未建立、资料发生变化、或建库 readiness 被阻断时使用。

    301 GitHub stars~5.6k tokensUpdated 9 days ago
    Documents & OfficeAuto-check passed
  • Cc Streaming Export Safety

    doccker/cc-use-exp

    当实现用户驱动的大文件导出或批量序列化(Excel/CSV/JSON/JSONL/PDF,数据量未知或超过 1 万行/10 MB)时触发;普通小文件下载、静态资源下载、非导出 Writer/Report 类不触发。防止 OOM、临时文件残留、同步导出阻塞 HTTP 线程和表格公式注入。

    1.1k GitHub stars~2.2k tokensUpdated 1 mo ago
    Documents & OfficeAuto-check passed

More from HKUDS/OpenSpace

All 199 skills in this repo
  • Walks through producing a master audio track plus stems in Python, from checking a reference file and timing sections by BPM to effects, a zip archive and final verification.

    7.7k GitHub stars~2.9k tokensUpdated 1 mo ago
    Auto-check passed
  • Handle cascading data retrieval tool failures by falling back to embedded knowledge generation

    7.7k GitHub stars~765 tokensUpdated 1 mo ago
    Auto-check passed
  • Gives an agent a workaround when its code-execution sandbox keeps failing: save the Python script to a file and run it through the shell instead.

    7.7k GitHub stars~588 tokensUpdated 1 mo ago
    Auto-check passed
  • A recovery routine for agents whose sandboxed code runner keeps failing: save the Python script to disk, then run it through the shell and read the output.

    7.7k GitHub stars~652 tokensUpdated 1 mo ago
    Auto-check passed
  • Fallback ladder for failed sandboxed code runs, plus the habit of fixing the working directory first so generated files land in the right place.

    7.7k GitHub stars~1.1k tokensUpdated 1 mo ago
    Auto-check passed
  • Fallback workflow for executing Python code when executecodesandbox fails repeatedly

    7.7k GitHub stars~1.1k tokensUpdated 1 mo ago
    Auto-check passed

Works with

Questions about Document Direct Python

What does Document Direct Python do?

Use direct Python execution for reliable document creation including spreadsheets, PDFs, and structured reports. Document Direct Python is an agent skill from HKUDS/OpenSpace.

When should I use Document Direct Python?

Document Direct Python fits situations like: tasks that involve Excel spreadsheets; tasks that involve PDF.

How do I install Document Direct Python in Claude Code?

Run `npx skills add HKUDS/OpenSpace --skill document-direct-python -a claude-code`. Or copy the skill folder (benchmarks/gdpval/skills/spreadsheet-direct-python-merged in HKUDS/OpenSpace) into .claude/skills/document-direct-python in your project. Claude Code loads it when a task matches its description.

How do I install Document Direct Python in Codex?

Run `npx skills add HKUDS/OpenSpace --skill document-direct-python -a codex`. Or copy the skill folder (benchmarks/gdpval/skills/spreadsheet-direct-python-merged in HKUDS/OpenSpace) into .agents/skills/document-direct-python in your project. Codex loads it when a task matches its description.

Can I use Document Direct Python in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add HKUDS/OpenSpace --skill document-direct-python -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/document-direct-python, .gemini/skills/document-direct-python, .github/skills/document-direct-python and .opencode/skills/document-direct-python in your project.

What does Document Direct Python need to run?

Going by SKILL.md and its folder, Document Direct Python needs the command-line tools its instructions call (python3). Our summary lists: Python 3.

Does Document Direct Python access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Document Direct Python safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Document Direct Python use?

Document Direct Python is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Document Direct Python use?

About 2.7k tokens (SKILL.md is roughly 11k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Document Direct Python?

Skills that share tags, products or a category with Document Direct Python: Instrument Data To Allotrope (aws-samples/amazon-bedrock-agents-healthcare-lifesciences, 274 stars), Doc Cleaner (notoriouslab/doc-cleaner, 309 stars), Mineru (Nebutra/MinerU-Skill, 122 stars) and Office To Md (shuyu-labs/WebCode, 278 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Document Direct Python?

HKUDS (a GitHub organization) maintains it in HKUDS/OpenSpace, which has 7,743 GitHub stars. The repository holds 199 skills in this directory. The repository was last updated on August 12, 2026.

Source: HKUDS/OpenSpace on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.