2.2K skills · sorted by relevance
Build image classification and search workflows by comparing images with text descriptions. Covers classification without additional training, similarity scoring, batch processing, and storing image representations in a vector database for retrieval.
Evaluate code generation models with HumanEval, MBPP, and other benchmarks. Test generated programs, compare models across programming languages, and record pass@k scores, which estimate the chance that at least one of k solutions passes the tests.
Find answers to employee questions about leave, benefits, expenses, and other company rules using supplied handbooks or connected knowledge bases. Receive plain-language explanations with source references, exceptions, and contacts for questions beyond the documentation.
Suggest or assess company, brand, and product names that reflect a founder's cultural roots and suit their target markets. Compare meanings, pronunciation, spelling, distinctiveness, and professional tone, with clear follow-up checks before committing to a name.
Review Python, JavaScript/TypeScript, and Go code for security issues using guidance for the project's languages and frameworks. Get prioritized findings, suggested improvements, or help writing new code with safer defaults.
Download YouTube captions and thumbnails from a URL or video ID. Save transcripts as Markdown or SRT subtitle files, select or translate languages, add chapters, and use AI processing to label speakers.
Set up the official Google Ads MCP Server, a connector that lets AI assistants retrieve campaign data and reporting metrics through natural language. The guide covers installation, credentials, client configuration, and connection checks.
Review Python, JavaScript/TypeScript, and Go code for security issues using guidance for the project's languages and frameworks. Get prioritized findings, suggested improvements, or help writing new code with safer defaults.
Render landscape videos with KrillinAI CLI, combining original footage with bilingual subtitles or prepared dubbed footage with target-language subtitles. Optionally replace audio, adjust subtitle styling, and verify the output video and subtitle placement.
Guide knowledge distillation, where a smaller language model learns from a larger teacher. Covers training from teacher responses or prediction scores, adjusting training settings, and comparing outputs to support model compression and specialized training.
Build image captioning, visual question answering, and text-based image search with BLIP-2. The skill guides model selection, batch processing, and memory optimization, with examples for understanding images without training for each specific task.
Train or fine-tune language and vision models on Hugging Face Jobs without a local GPU. Covers dataset checks, hardware and cost planning, training monitoring, Hub storage, and GGUF conversion for local use.
Turn legal clinic case notes into status update drafts for clients, supervising professors, or courts. Adjust language, detail, and format for each reader, covering recent progress, upcoming events, outstanding issues, and next steps as appropriate.
Find the replacement command for plain-language client correspondence. This deprecated reference directs routine letters to `/client-letter` and substantive status updates to `/status client`, helping users choose the appropriate workflow when moving from the older skill.
Convert text into spoken audio for video narration, audiobooks, reading assistance, or language practice. The skill uses Edge-TTS preset voices in languages including Japanese, English, and Chinese, with voice selection guided by the text and requested tone.
Get a quick structural overview of a codebase with Trailmark. The skill identifies programming languages, counts entry points where execution begins, and returns dependency information to support orientation before deeper analysis.
Review changes or finished work, check claims against tests, logs, running software, or source files, and explain the findings in plain language. It separates verified facts from assumptions and makes incomplete checks and remaining uncertainty clear.
Change DeepChat preferences through conversation, including theme, interface language, font size, sound effects, and whether copied text includes COT details. Dedicated tools apply supported changes, while other requests open the relevant settings section with guidance.
Extract Amazon search results with product titles, prices, ratings, identifiers, and available specifications. Set keywords, brand, page count, and browsing language to gather structured information for market research, competitor comparisons, and product catalog updates.
Provide language-specific reference files for test analysis skills. The catalog helps agents recognize test methods, assertions, skipped tests, waiting patterns, setup and cleanup hooks, external dependencies, and tagging capabilities across different frameworks.
Turn plain-language questions into SQL for SQLite and PostgreSQL using database structure as context. Review queries against 13 optimization rules and interpret execution plans, which show how the database retrieves data, to investigate slow queries.
Set up and audit multilingual application interfaces. Replace hard-coded text with translation keys, localize error messages, and check for missing translations, matching keys across languages, plural forms, and date or number formatting.
Investigate product demand through Reddit, social posts, transcripts, and comments. Get sourced customer pain points, buying signals, existing alternatives, and objections, plus messaging angles and product tests grounded in the language people actually use.
Extract colors, fonts, spacing, and other design choices from a website URL. Generate eight files for review and reuse, including framework themes, CSS variables, Figma variables, and an accessibility score against WCAG guidelines.