Recently updated open-source projects

html-deck treats HTML as the editable source for fixed 1280×720 slides, renders each section in a local Chromium-based browser, and assembles the screenshots into a 16:9 PPTX or a folder of PNGs. It suits dense, consistent presentations with code, charts, tables, formulas, or CJK text. The important trade-off is that the PowerPoint contains one full-slide bitmap per page, so text is not selectable there; Node 18 and a Chromium browser are required.

Use libu-studio to track prompts, generation runs, candidates, costs, character profiles, and delivery rules for liblib.art assets in one SQLite database and CLI. The workflow supports icons, images, videos, and WebP animations, including optional injection into a target project. Foreground masking and despill require macOS 14+; generation also needs a liblib account and credits, while other systems are limited to the data model and scripts.

teach-me-a-lesson helps complete beginners understand Claude, AI tools, and programming through one clear analogy at a time. It combines explanation with blame-free troubleshooting, can save knowledge notes, flag wasted context, and adapt its language and tone. While you work, it warns about risky commands, exposed keys, and irreversible actions without deciding for you. Personal knowledge files are optional.

wp-plugin-skill turns a natural-language plugin description into WordPress code, an installable ZIP, an update mechanism, and a landing page. Its generation process uses a plugin blueprint covering admin, front end, database, REST, scheduled work, security, and distribution, together with templates and a quality checklist. It is aimed at agencies, merchants, and others who need custom PHP without writing it themselves. The generated package still needs testing and installation in the target WordPress environment.

github-growth-check turns GitHub traffic metadata into a diagnostic brief that tells you whether a repository is growing, flat, or simply past its launch spike—and what to do next. It compares traffic, clones, referrers, popular paths, and stars using source-specific benchmarks, excludes the API’s latest reporting lag, and stores snapshots for longer history. It is read-only and requires an authenticated GitHub CLI with access to the target repository.

claude-github-launch turns a trigger phrase into a guided GitHub launch: it checks authentication and secrets, reads the project, creates plugin and CI structure, generates a social preview, writes a landing-page README, and adds changelog, license, and ignore files. It can then set topics, labels, Discussions, and a release, followed by ready-to-paste copy for Reddit, X, Hacker News, and an awesome list. Existing skill content and environment files are protected, and a secret scan stops the process when needed.

Chiron accepts repositories, PDFs, vocabulary CSVs, and research papers, then produces a self-contained interactive lesson HTML. It supports separate learning paths for code, medicine, and language, using a five-stage ingest, generate, validate, build, and assemble pipeline with structured-output retries. Each lesson has its own SQLite state, while the parent Claude Code agent handles the text-LLM steps.

memphis-html-ppt turns local Markdown or a web page into a self-contained HTML slide deck with bold color, geometric accents, and editorial layouts. It extracts and cleans the source, plans slide content, renders the preview, and checks structure, content quality, and template fit before sharing. The output opens in a browser, while complex article structure and long sections may still be compressed.

DeepArchi frames AI agents as organizational roles and supplies a nine-agent operating model with governance and task-orchestration topology. It adds cross-model validation, a human charter, and explicit security constraints, plus ArchiMate 3.2 guidance with Draw.io integration. It is aimed at teams designing structured multi-agent organizations rather than a general-purpose agent runtime.

BIAN Skill helps banking architecture teams describe capabilities through standardized business areas, service domains, and service operations, then use those definitions to shape APIs and system boundaries. It covers domains such as customer and party data, payments, products, channels, risk, and compliance, with guidance for current-state mapping, gap analysis, and phased modernization.

This skill organizes the TOGAF ADM method, content framework, and governance practices into a working reference for architects and AI agents. It helps plan activities and deliverables from architecture vision through business, data, application, technology, migration, and change management, while also covering viewpoints, architecture principles, and ABB-to-SBB modeling.

things-app connects Claude to Things 3 with AppleScript so it can read lists and create, update, move, reschedule, complete, or delete tasks and projects. It also imports a Reminders list as a flat Things project and turns YAML templates into reusable projects with sections. The skill requires Claude Code and Things 3 on the same macOS machine, so it is not suitable for Cowork’s remote Linux shell. Reminders imports intentionally lose section structure.

DDD Skill guides an AI agent through strategic and tactical Domain-Driven Design. It helps identify subdomains, establish a ubiquitous language, split bounded contexts, and design entities, value objects, aggregates, domain services, and events. It also covers architectural patterns such as hexagonal architecture, CQRS, event sourcing, and Saga for complex systems and microservice boundary decisions.

This OpenSkills package guides an AI agent to create or revise Draw.io-format diagrams, including flowcharts, architecture diagrams, network topologies, UML, ER diagrams, and mind maps. The generated files can be opened and refined in Draw.io after creation, making the skill useful for turning written descriptions into structured diagrams.

The review skill inspects Java 8, 11, or 17 code for issues such as null handling, monetary precision, thread safety, dated APIs, and concurrency choices, then writes an analysis file with priorities and a 1–100 score. After approval, the implementation skill applies the recommendations to a new *Modern.java file beside the original. Its rules are especially tailored to financial and messaging systems.

This skill turns clinical .xlsx or .csv data into a standalone Korean HTML exploratory-analysis dashboard with dataset summaries, missingness, distributions, outlier candidates, optional Table 1, correlations, and VIF. It embeds assets locally, masks identifier columns, and supports print or PDF output, but does not perform inferential analysis or handle images, waveforms, or free text.

Run CelesteOps on macOS to track tasks, content dates, milestones, documents, comments, and sandboxed HTML prototypes in one local-first workspace. Its MCP shim exposes the app to Claude, Cursor, Codex, and other clients, while the public repository distributes the built app and connection kit rather than the app source.

generate-presentation teaches Claude to choose a deck-level narrative such as the Pyramid Principle, SCQA, AIDA, or a problem-solving flow, then compose individual slides with reusable patterns. The guided process collects the topic, audience, and tone before producing a Markdown outline that can be moved into PowerPoint, Keynote, Gamma, or another presentation tool. It is a skill specification, not a standalone slide-rendering application.