Voice profiles, positions, and philosophy of 66 Drupal core contributors — extracted from real issue queue activity, not summaries.
Built from 53,102+ issue queue comments, 14,042 commits,
and 20+ years of Drupal history
A research infrastructure for understanding how Drupal core is shaped — the people, positions, and philosophies behind 20+ years of architectural decisions.
Extracted positions, terminology, and stylometrics from real issue queue activity. Not summaries — pattern models trained on thousands of comments per author, with fidelity gates and adversarial validation.
Who reviews whom on what. Where the fault lines are. Which contributors align on caching but split on API design. A philosophy graph across 40 topics, 1,711 author pairs, mined from 68K+ commits.
Install and ask "What would catch say about X?" or "How would xjm and Crell differ on API design?" for any of the 66 contributors. Query, review, and compare skills — all installable via one command.
Gold tier: 2,000+ comments. Silver: 500+. Bronze: 200+. Profiles are stratified by data volume — richer evidence, higher fidelity.
Showing 66 contributors across all tiers
Data-driven tools for patch authors, reviewers, and community researchers. All tools read from the local corpus — no API calls needed.
All 143 skills install in one command. Ask any contributor a question directly from your terminal.
# Clone and install all 143 skills
git clone https://github.com/zivtech/drupal-core-bot
bash install-skills.sh
# Then ask any contributor about any topic
claude "What would catch say about entity caching?"
claude "How would xjm and Crell differ on API design?"
claude "Simulate a core team review of this patch"All data sourced from public drupal.org issue queues (CC BY-SA 2.0). Every corpus entry tracks its original URL, comment ID, issue ID, and date. Master index at corpus/provenance-index.json.
Profiles are research instruments, not authoritative statements about any individual. They model patterns in public writing — they cannot substitute for direct communication with contributors.
All synthesized summaries are AI-generated from evidence — not direct quotes. Profile outputs represent observed patterns in the corpus, with fidelity scores and confidence intervals documented.
Any contributor can request removal of their profile. Open an issue on GitHub with the subject "Profile removal request" — we will process it within 5 business days without requiring justification.
The most significant gap in the corpus is Drupal Slack, where a substantial portion of core development discussion happens informally. Architecture debates, review coordination, mentoring, and real-time decision-making all occur in Slack channels that are not publicly accessible or archivable. Profiles are built exclusively from drupal.org issue queues, git commits, change records, and blog posts. For contributors whose primary engagement is through Slack, the profiles may underrepresent their actual influence and positions.
If Drupal Slack data were accessible through community-sanctioned exports or a public archive, it would improve position accuracy, debate coverage, relationship mapping, and mentoring pattern detection.
The current pipeline routes queries to relevant experts but cannot perform full-text search across the 853 MB corpus. Questions like "what does contributor X specifically think about topic Y?" require search-indexed infrastructure (e.g., Solr) to enable direct evidence retrieval from the 300K+ comments.