evo-pdf-content-extraction

SkillDocs & knowledge

Extracts and analyzes text content from PDF pages to identify author-revealing information including names, affiliations, emails, arXiv IDs, venue/conference mentions, and other identity-leaking content.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the evo-pdf-content-extraction skill

What this skill tells your AI

The instructions your AI receives, as published by openlair/openskill in tasks-evolved/paper-anonymizer/environment/skills/evo-pdf-content-extraction/SKILL.md and read by ahel’s review.

Extracts text from academic PDFs and identifies all content that could reveal authorship for anonymization purposes.

Key Functions

  • extract_page_text(page) - Get plain text from a page
  • extract_words_with_positions(page) - Get words with bounding box coordinates
  • identify_author_block_region(page) - Find the author block area on first page
  • detect_arxiv_identifiers(text) - Find arXiv IDs in text
  • detect_email_addresses(text) - Find email addresses
  • detect_affiliations(text) - Find institutional affiliations
  • detect_venue_statements(text) - Find venue/conference acceptance statements
  • get_all_redaction_targets(pdf_path) - Main function: analyze a PDF and return all strings to redact

Usage

import sys
sys.path.insert(0, '/app/environment/skills/evo-pdf-content-extraction/scripts')
from extraction import get_all_redaction_targets

targets = get_all_redaction_targets('/root/paper1.pdf')
print(targets['exact_strings'])  # List of exact strings to redact
print(targets['regex_patterns'])  # List of regex patterns to apply
print(targets['link_patterns'])   # URL patterns to scrub

Signals

GitHub stars
89
Forks
4
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
evo-pdf-content-extraction
Source
github.com/openlair/openskill
evo-pdf-content-extraction: Skill · ahel