-
Notifications
You must be signed in to change notification settings - Fork 0
Getting Started
This page explains how to install and use skills from this repository in every Claude interface.
No prerequisites are needed to use skills that are pure instructions (no scripts). For skills that include Python or shell scripts you will need:
| Dependency | Required by | Install |
|---|---|---|
| Python 3.8+ | docx, pdf, pptx, xlsx, skill-creator, webapp-testing, slack-gif-creator | System package manager |
| Node.js & npm | web-artifacts-builder, docx (JS library) | https://nodejs.org |
| LibreOffice | docx, pptx, xlsx (format conversion) | sudo apt install libreoffice |
| Poppler utils | pdf, pptx (PDF โ image) | sudo apt install poppler-utils |
| Playwright | webapp-testing | pip install playwright && playwright install chromium |
| Tesseract OCR | pdf (OCR mode) | sudo apt install tesseract-ocr |
Most users only need Python 3.8+; install extra dependencies only for the skills you intend to run scripts from.
Claude Code is the recommended environment for local development workflows that involve skill scripts.
/plugin marketplace add anthropics/skills
/plugin install document-skills@anthropic-agent-skills
/plugin install example-skills@anthropic-agent-skills
/plugin install claude-api@anthropic-agent-skills
| Plugin group | Skills included |
|---|---|
document-skills |
docx, pdf, pptx, xlsx |
example-skills |
algorithmic-art, brand-guidelines, canvas-design, doc-coauthoring, frontend-design, internal-comms, mcp-builder, skill-creator, slack-gif-creator, theme-factory, web-artifacts-builder, webapp-testing |
claude-api |
claude-api |
Just mention the skill in your request:
Use the PDF skill to extract the form fields from /path/to/form.pdf
Use the DOCX skill to create a project proposal document
Use the algorithmic-art skill to create a flow field composition
Claude will automatically load the matching skill and follow its instructions.
Clone this repository and point Claude Code at the skill folder directly:
git clone https://github.com/d33disc/upstream_skillsThen in Claude Code:
/skill load ./upstream_skills/skills/mcp-builder
All skills in this repository are already built into paid Claude.ai plans โ no installation required.
To upload your own custom skill or to activate a specific skill:
- Open a conversation in Claude.ai.
- Click Skills in the left sidebar.
- Click + Add skill and upload your
SKILL.md(and optional supporting files). - Enable the skill for the current project.
See the official guide: Using skills in Claude.
Anthropic's document skills (docx, pdf, pptx, xlsx) are available as built-in API skills. Refer to the Skills API Quickstart for full documentation.
import anthropic
client = anthropic.Anthropic()
# 1. Upload the skill file(s)
with open("skills/mcp-builder/SKILL.md", "rb") as f:
skill_file = client.beta.files.upload(
file=("SKILL.md", f, "text/markdown"),
)
# 2. Reference the skill in a message
response = client.beta.messages.create(
model="claude-opus-4-5",
max_tokens=4096,
messages=[{"role": "user", "content": "Build an MCP server for the GitHub API."}],
system=[{"type": "tool_result", "tool_use_id": skill_file.id}],
betas=["skills-2025-03-01"],
)
print(response.content[0].text)Note: API feature flags and endpoint paths may change. Always consult the official API docs for the latest.
Some skills include Python utility scripts that you can run directly outside Claude. Below are common entry points.
# Unpack a Word document for XML editing
python skills/docx/scripts/office/unpack.py document.docx unpacked/
# Repack after editing
python skills/docx/scripts/office/pack.py unpacked/ output.docx --original document.docx
# Validate the result
python skills/docx/scripts/office/validate.py output.docx
# Recalculate Excel formulas with LibreOffice
python skills/xlsx/scripts/recalc.py workbook.xlsx# Extract fillable form field metadata
python skills/pdf/scripts/extract_form_field_info.py form.pdf
# Fill form fields
python skills/pdf/scripts/fill_fillable_fields.py form.pdf filled.pdf field_values.json
# Convert PDF pages to images (for OCR or review)
python skills/pdf/scripts/convert_pdf_to_images.py doc.pdf output_dir/# Scaffold a new React + Tailwind + shadcn/ui project
bash skills/web-artifacts-builder/scripts/init-artifact.sh my-app
cd my-app && npm run dev
# Bundle to a single self-contained HTML file
bash skills/web-artifacts-builder/scripts/bundle-artifact.sh# Run Playwright automation against a local dev server
python skills/webapp-testing/scripts/with_server.py \
--server "npm run dev" --port 5173 \
-- python my_automation.pyfrom skills.slack_gif_creator.core.gif_builder import GIFBuilder
from PIL import Image, ImageDraw
builder = GIFBuilder(width=128, height=128, fps=10)
for i in range(12):
frame = Image.new("RGB", (128, 128), (240, 248, 255))
draw = ImageDraw.Draw(frame)
draw.ellipse([(i * 8, i * 8), (64 + i * 4, 64 + i * 4)], fill=(70, 130, 180))
builder.add_frame(frame)
builder.save("output.gif", num_colors=48, optimize_for_emoji=True)| Symptom | Likely cause | Fix |
|---|---|---|
ModuleNotFoundError: No module named 'pypdf' |
Python dependency missing | pip install pypdf |
soffice: command not found |
LibreOffice not installed | sudo apt install libreoffice |
pdftoppm: command not found |
Poppler not installed | sudo apt install poppler-utils |
| Script exits with LibreOffice lock error | Previous LibreOffice instance still running | pkill soffice |
| Claude does not trigger the skill | Skill not installed / not enabled | Re-install or enable in the Skills panel |
| Generated file fails validation | XML structure error | Run validate.py, inspect the diff |
See also: Architecture ยท Contributing ยท Skills Overview
Wiki auto-generated from https://github.com/d33disc/upstream_skills. Last updated: 2026-03-30 18:33 UTC.
๐ Home
๐ Skills Overview
๐ Documentation
๐จ Creative & Design
โ๏ธ Development & Technical
๐ข Enterprise & Communication
๐ Document Skills
๐ง Other