Recently updated open-source projects

The skill gathers names, date, venue, languages, photos, and preferred visual directions through conversation, then creates a custom HTML invitation for each selected direction. A local tweak studio lets you adjust colors, fonts, photo frames, and optional elements before exporting social and print-sized PNG files. It requires Node.js 18 and a Chromium-family browser, while the couple's photos and personal details remain on the local machine.

frontend-skills helps an agent build or refactor React, Next.js, and Tailwind applications by deciding where code should run before choosing components, actions, or route handlers. It covers Next.js 16 caching, composition patterns, semantic Tailwind themes, common UI shapes, accessibility, and high-impact performance mistakes. The skill is aimed at teams that want fast interfaces without making future changes expensive.

This Claude Code skill scaffolds an audit table, activity page, API route, and logging helper for a Next.js and Supabase admin dashboard. Developers wire the helper into successful admin actions and can add manual entries for deployments or operational work. RLS protects client access and blocks updates and deletions, while the server helper assumes an already admin-gated route. The templates expect a specific App Router, auth, UI, and Supabase structure, and logging failures do not block the underlying action.

Connect the script to an HTML demo and your own Supabase project so reviewers can pin comments to exact positions, reply, change resolution states, and jump between comments across pages. It also offers live synchronization and Markdown export. There is no account or permission system, so anyone with the link can comment or delete entries; it is intended for prototype review rather than production data.

Search popular Xiaohongshu notes by keyword for topic ideas, then publish an image-and-text post with a title, body, images, tags, and visibility settings. Posts can be scheduled from one hour to fourteen days ahead, and QR-code login saves the session for reuse or switching accounts. Python 3.8 or later and the xiaohongshu-mcp binary are required.

claude-skill-status gives a Claude Code session a terse `/status` snapshot of shipped work, commits, pending user actions, and the next step. It reads Todo state, Git history, and the working tree, but does not write files or maintain project-level planning. Use it when switching context or returning to a long session and needing a quick orientation.

Audit a LaTeX paper before submission for anonymity, reference hygiene, required sections, and page limits measured from a real PDF compile. A reviewer-style checklist also flags issues such as missing model versioning or mismatched ablations. Venue rules arrive through overlay skills; this is an audit toolkit, not a drafting assistant.

Pitch Generator asks for missing context such as audience, goal, company stage, traction, timing, and the requested next step before producing a one-line pitch, a short paragraph, and an investment-style version. For pre-seed fundraising, it emphasizes problem validation, founder-market fit, and use of funds while adapting to different investor types. Local benchmark decks can inform messaging patterns without copying their claims or wording.

This Claude skill is for designers working in the Elementor Pro V4 visual editor rather than PHP addon developers. It uses the correct panel field names and a class-first decision hierarchy to guide layouts, variables, dark mode, four responsive breakpoints, native interactions, and troubleshooting. It targets Elementor Pro 4.x on WordPress 6.x.

ZAFU Thesis Formatter converts DOCX or Markdown theses into a Zhejiang A&F University LaTeX/PDF template. A Python script generates the cover, abstracts, table of contents, chapters, acknowledgements, and bibliography. It requires Python and a TeX distribution, and the included format targets ZAFU.

Pressure-test a structured draft against three readers whose goals and expectations deliberately differ. The skill runs their reviews in parallel and consolidates the results into consensus issues, author tradeoffs, and useful outlier flags, rather than averaging disagreement away. It is intended for editorial revision, not copyediting, factual verification, code review, or early skeleton drafts.

ads-cite lets astronomy and astrophysics authors search NASA ADS from Claude Code or a standalone CLI using authors, years, titles, DOI, arXiv IDs, and other fields. After a result is selected, it fetches the verbatim ADS BibTeX entry and appends it to a .bib file with duplicate checks; it can also list citing and referenced papers or generate memorable citekeys. A NASA ADS API token and Python 3.9+ are required.

say-less changes Claude's framing for designers: lead with the user-facing behavior and design implication, keep the English technical term as an anchor, and explain the mechanism only as needed. It covers frontend web, mobile, and AI/LLM/agent topics through on-demand references, adapts prose to the user's working language, and preserves terms such as hydration, RAG, and MCP.

Answer up to four questions about the learning objective, subject, audience knowledge, and constraints. The skill then selects an appropriate diagram or illustration approach, explains the instructional reasoning, recommends Mermaid, PlantUML, or an image model, and returns a prompt you can paste into that tool. It is intended for technical course diagrams, illustrations, and infographic panels, not data charts.

This skill turns an interest description into multiple GitHub searches, merges and deduplicates the results, then ranks them by relevance, recent activity, stars, or trending velocity. It supports filters, saved profiles, caches, and Markdown or JSON output, but requires an authenticated gh CLI and remains read-only.

Choose a fictional character and source work, then guide the skill through source research, trait extraction, and construction of an activatable character profile in Claude Code. It captures core conflicts, behavior patterns, relationship dynamics, speech style, and performance boundaries, followed by dialogue, scenario, and boundary tests to validate the result.

Give it an MP3, WAV, FLAC, M4A, or OGG file to separate vocals, transcribe lyrics, and extract musical features such as BPM, key, and spectral characteristics. An AI agent then turns the raw results into genre, mood, instrument, vocal, and production analysis. Python and ffmpeg are required; a CUDA GPU is recommended but not mandatory.

This skill evaluates AI pilots, vendor proposals, and investment theses through five perspectives: investigator, devil’s advocate, long-term strategist, realist, and senior advisor. It tests the problem, operating mode, durability, data and integration readiness, then returns a verdict such as Fund-with-condition, Reframe, or Kill plus one concrete action.