mr-devs

Use when the user wants to convert a project's `paper/` directory into a git submodule pointing at a separate paper repo (typically for Overleaf linking). Handles removing any existing tracked `paper/` content, adding the submodule, writing a standardized README inside the submodule, and updating the parent repo's README.

Clean up PDF filenames by removing spaces, em dashes, and special characters

Create BibTeX entries from PDF papers and add to main.bib

Create a new Python script following project best practices. $ARGUMENTS

Create markdown summaries for PDF papers (creates BibTeX first if needed)

Clean README.md files in the entire repository or specified directories. $ARGUMENTS

Use when updating or replacing the writing style configuration for the paper-writing pipeline, targeting a new venue, audience, or style preferences.

Write LaTeX paper sections based on outline.md using citations from main.bib

Generate a cross-cutting synthesis document from all summaries in summaries/. Reads synthesis-guidance.md if present. Can trigger the full pipeline if summaries are missing.

Package the synthesis into a shareable ZIP for distribution. Use when the user wants to share, export, or distribute their synthesis to others. Creates a self-contained directory (synthesis-export/) with the HTML, CSS, JS, all source PDFs, and all summary files, then zips it. The "Ask Claude" functionality is automatically removed (button hidden via CSS) since it requires a local server to work. Invoke when the user says things like "export", "share", "package", "zip up", "distribute", or "send the synthesis to someone".

Use when the user wants to view or interact with synthesis.html after running /create-synthesis — builds the HTML, verifies Claude Code auth, starts the local server, and opens the page in the browser

Reset the cc-synthesizer pipeline by deleting all generated files so a fresh batch of PDFs can be processed from scratch. Invoke whenever the user wants to start over with new documents, clear the workspace, or clean up pipeline outputs. Trigger on phrases like "reset", "clean up", "start fresh", "new batch", "clear the pipeline", "delete generated files", "wipe the outputs", or "prepare for new PDFs". Always checks for an existing export first before deleting anything — don't skip this check.

Generate per-document markdown summaries from PDFs. Detects document type and adapts the summary structure accordingly. Updates citations.json.

Clean up PDF filenames by removing spaces, em dashes, and special characters

Use when adding a talk, poster, invited talk, or demonstration to Matthew DeVerna's LaTeX CV bibliography at /Users/mdeverna/Documents/Projects/cv/ref.bib

Use when adding a publication (journal article, conference paper, preprint, manuscript under review, dissertation, blog post) to Matthew DeVerna's LaTeX CV bibliography at /Users/mdeverna/Documents/Projects/cv/ref.bib

Use when adding or editing entries in Matthew DeVerna's LaTeX CV section files (education, awards, teaching, service, advising, memberships, media, tools, other experience)

Serves the academic website locally and verifies it renders correctly — all JSON data files parse and every section renders from its data. Use to preview the site, check changes after editing data or design files, or debug a section that isn't showing up.

Designs and builds the academic website from design references the user provides — reference URLs, pasted screenshots/images, or free-form design notes. Use for the initial setup of this template, a full redesign, or significant restyling. Restyles the existing JSON-driven template rather than generating a site from scratch.

Updates the data/*.json content files that drive the academic website — working papers, published papers, news items, talks, software, ongoing projects, teaching, and the profile/bio. Use when adding a preprint or paper, converting a working paper to published, or adding/editing news, talk, software, project, teaching, or profile entries.

Search for and update conference deadline data in deadlines.json for tracked conferences

Use when verifying that upcoming deadlines in the Deadline Hub project are accurate. Fetches source webpages and generates a markdown report with evidence and justification for each deadline date.

Use when the user wants to create a new Claude Code skill (global or project-specific). Guides the full skill authoring process including naming, structure, frontmatter, and file creation.

Manuscript and pre-analysis plan screening assistant for political science, communication, and information science

Use when the user wants to release, publish, or ship a new version of fetchbib to PyPI, cut a release, or bump the package version.