Notre avis
Cette compétence permet de lire, extraire et analyser le contenu de documents PDF via des commandes simples.
Points forts
- Lecture native des PDF sans outil externe
- Extraction ciblée d'informations (tableaux, citations, contacts)
- Capacité de résumé et de comparaison entre plusieurs PDF
Limites
- Les PDF scannés sans OCR ne peuvent pas être extraits correctement
- Certains PDF protégés empêchent l'extraction de texte
- Les très gros PDF (100+ pages) peuvent nécessiter une pagination manuelle
Utilisez cette compétence lorsque vous devez extraire ou résumer le contenu textuel d'un fichier PDF.
Évitez de l'utiliser pour des PDF contenant uniquement des images scannées sans texte sous-jacent, car l'OCR n'est pas intégré.
Analyse de sécurité
PrudenceThe skill primarily uses the built-in Read tool for PDF files, which is low-risk. However, it also instructs users to run a Python utility via `uv run`, a command that executes a Python script. While the script likely performs benign PDF operations, any command-line execution carries potential risk if the tool is maliciously modified or input is not sanitized. No destructive or exfiltration commands are directly present.
- •Uses a bash command `uv run tools/pdf_util.py` which invokes a Python script; if the script or its dependencies are compromised, it could lead to arbitrary code execution. However, the script is local and intended for PDF processing, so risk is moderate.
Exemples
Read /path/to/document.pdf and give me a concise summary of the main points.Extract all tables from /path/to/report.pdf and display them as markdown.Compare /path/to/draft1.pdf and /path/to/draft2.pdf and list the differences.name: pdf description: This skill should be used when the user asks to "read pdf", "view pdf", "extract text from pdf", "summarize pdf", or shares a PDF file path.
Read, extract, and analyze PDF documents.
Instructions
When the user provides a PDF path or asks about PDF content:
1. Reading PDFs
Use the Read tool directly on the PDF file. Claude Code supports reading PDF files natively:
Read: /path/to/document.pdf
This extracts both text and visual content for analysis.
2. Common Tasks
Summarize: Provide a concise summary of the document
- Key points and main arguments
- Document structure overview
- Important figures, tables, or data
Extract specific info: Find and extract:
- Tables (convert to markdown format)
- Code snippets
- Citations/references
- Contact information
- Dates and deadlines
Compare PDFs: When given multiple PDFs:
- Identify differences
- Highlight common themes
- Cross-reference information
3. Python Utility
For batch operations or programmatic access:
uv run tools/pdf_util.py extract /path/to/file.pdf
uv run tools/pdf_util.py info /path/to/file.pdf
uv run tools/pdf_util.py search /path/to/file.pdf "search term"
4. Output Format
When presenting PDF content:
- Use markdown headers for document sections
- Convert tables to markdown tables
- Note page numbers for reference:
(p. 5) - Flag any extraction issues (scanned images, encrypted content)
5. Limitations
- Scanned PDFs may have limited text extraction (OCR not included)
- Some PDFs have copy-protection that prevents text extraction
- Very large PDFs (100+ pages) may need pagination
Priorisation de Tâches
Productivite
Priorise vos tâches avec les frameworks Eisenhower, ICE et RICE.
Generateur de Rapport Hebdomadaire
Productivite
Generez des rapports de statut hebdomadaires structures et concis.
Rapport de Daily Standup
Productivite
Génère des rapports de daily standup structurés et concis.