HKU-MMLab
Official@HKU-MMLab · Hong Kong
Multimedia Lab at The University of Hong Kong, a top-notch research cohort dedicated to research and educating the best minds.
Agent Skills by HKU-MMLab
Showing 78 vetted skills indexed across 1 GitHub repositories.
HKU-MMLabapt-package-manager
Search, install, and verify Debian or Ubuntu packages with apt-get in containers.
HKU-MMLabagent-browser-control
Controls a headed Chromium browser via the agent-browser CLI for web interaction and screenshots.
HKU-MMLabduckduckgo-search
Search the web via DuckDuckGo endpoints without requiring an API key.
HKU-MMLabdesktop-control
Automates mouse, keyboard, screen capture, and window management operations on desktop environments.
HKU-MMLabweb-search
Search web pages, news, images, and videos via the DuckDuckGo API.
HKU-MMLablinux-gui-control
Automate X11 desktop applications using xdotool, wmctrl, and the AT-SPI accessibility tree.
HKU-MMLablog-analyzer
Parse, search, and analyze application logs across plain text and JSON formats.
HKU-MMLabmarkdown-formatter
Format, lint, and beautify markdown documents using configurable style guides.
HKU-MMLabTranslate
Translate text across languages while preserving formatting, placeholders, and locale conventions.
HKU-MMLabFFmpeg Video Editor
Generates FFmpeg commands from natural language video editing requests.
HKU-MMLabopenai-whisper-api
Transcribe audio and video files via OpenAI-compatible Whisper transcription APIs.
HKU-MMLabpymupdf-pdf
Parse PDF files into Markdown and JSON with optional image and table extraction using PyMuPDF.
HKU-MMLabxlsx-cn
Create, edit, and recalculate Excel spreadsheets with formulas, formatting, and error validation.
HKU-MMLabdocument-pro
Extract text, tables, and key information from PDF, DOCX, PPTX, and XLSX documents.
HKU-MMLabcsv-pipeline
Process, transform, join, and report on CSV, TSV, and JSON tabular data.
HKU-MMLabpython-dataviz
Generate static and interactive charts from data using matplotlib, seaborn, and plotly.
HKU-MMLabDocker
Diagnose and harden Docker containers, images, Compose stacks, networking, and volumes.
HKU-MMLabsecurity-auditor
Reviews code for OWASP Top 10 vulnerabilities and provides structured security audit reports.
HKU-MMLabnewsletter-blog-writer
Drafts newsletter blog posts from local data analysis with cited metrics.
HKU-MMLabcode-analysis
Analyze Git repositories and generate developer evaluation reports in Markdown, HTML, JSON, and PDF.
HKU-MMLabtest-runner
Parses test results from CI logs and reports to categorize failures and summarize suite status.
HKU-MMLabwho-is-actor
Analyzes Git commit history to profile developers across six behavioral dimensions.
HKU-MMLabpdf-smart-tool-cn
Convert, merge, split, encrypt, and OCR PDF documents through guided workflows.
HKU-MMLabtender-finance-local
Analyzes tender pricing, spreadsheet totals, and vendor financial tradeoffs from local files.
Frequently Asked Questions About HKU-MMLab
FAQPage SchemaWhat tasks can I accomplish with HKU-MMLab's skill registry?▼
The 80 skills cover PDF/DOCX/XLSX creation and parsing, CSV/JSON data pipelines, SQL querying across SQLite/PostgreSQL/MySQL, security audits and penetration testing, Git and GitHub operations, Docker management, web search, desktop GUI control, FFmpeg video editing, Whisper transcription, and stock analysis.
Who are these skills designed for?▼
Developers, sysadmins, security reviewers, data analysts, and office productivity users. Specialist skills target database engineers (query optimization, migrations), security auditors (OWASP, SAST, pentest), and business users (Excel dashboards, tender pricing, newsletter drafting, translation).
What are the installation and runtime requirements?▼
Requirements vary per skill and are declared in frontmatter: common binaries include python3, git, docker, ffmpeg, tesseract, pdftotext, and node. Some skills need environment tokens such as NOTION_API_TOKEN, TRELLO_API_KEY, AIRTABLE_PAT, or WHISPER_API_KEY; many run fully local with no network calls.
Are these skills open source and free to use?▼
Most skills carry permissive licenses — MIT or MIT-0 are explicitly declared on skills like test-runner, discord, code-analysis, and excel-weekly-dashboard. One exception is the comprehensive 'pdf' skill, which is proprietary with terms in its LICENSE.txt file.
Do the skills work offline without external services?▼
Many do: ocr-local, summarize-pro, pdf-ocr-local, and diagram-local-mermaid explicitly process data locally with no external requests. Others require credentials for hosted services, including Slack, Discord, Trello, Todoist, Google Calendar via gcalcli, Zoom, and OpenAI Whisper transcription.