Loading...
Loading...
Found 977 Skills
Local vision CLIs: glance (describe/ask/OCR an image), ground (locate a target, pixel box), detect (element inventory), trace (image to SVG geometry), crop (cut a pixel box to a file), and scripts/html_shot.py (HTML file to image). Use for any task involving an image — questions, text, splitting and transcribing long screenshots or chat histories, locating elements, comparing, rebuilding as HTML/SVG, digitizing a sketch or diagram, reading values off a chart, operating a GUI from screenshots — and to re-check an image yourself when a description you were given lacks a detail.
Create basic Transformers model adapters for msModelSlim. Implements required interfaces and completes a four-step verification workflow: generate test model -> full fallback quantization -> weight verification -> quantization description validation. Use this skill when the user wants to: (1) create msModelSlim adapters for decoder-only LLM, (2) adapt understanding VLM text backbones for quantization, (3) implement W8A8/W4A16 quantization workflow for new models. Trigger: user mentions "msModelSlim", "adapter", "model adapter","quantization", "W8A8","W4A16", "transformers", "LLM", "VLM", "adapter creation", "适配器","模型适配", "量化", "模型适配器", "LLM量化"
Hunts complexity in the current diff and APPLIES the simplifications directly to the working tree, then verifies the build. Five angles: reuse (duplicate logic, hand-rolled stdlib, platform features), quality (React/TypeScript hygiene, over-memoisation, exhaustive-deps, `any`, `CLAUDE.md`/`AGENTS.md` violations), efficiency (unnecessary work, missed concurrency, hot-path bloat), altitude (bandaid fixes special-cased onto shared code), and test discipline (repro test per bug fix, useless tests deleted, new tests only when they prevent a named failure). Hand it a `pr-reviewer` report and it applies those confirmed findings too. Use when the user says "tidy this up", "simplify", "clean up this diff", "polish my changes", "make this simpler", "apply the review findings", or "any reuse opportunities?". For a read-only report instead, use `pr-reviewer`; for the PR's title, description, or commits, use `pr-creator`.
Find shadcn registry items by keyword to install, compare registries, or verify before add. Use when the user wants a component but didn't give @registry/name or asks which registry has X — even without mentioning MCP. Keyword on names/descriptions only; not layouts, blocks, or semantic search (shoogle.dev/search). Skip known add targets.
The single entry point into the daily-work arc — a hybrid menu/argument router. Trigger whenever the user invokes /dev-workflows:daily, types /daily, asks "what should I do now", "where was I", "start my day", "what's next today", "wrap up my day", "end my day", or seems unsure which skill fits their current daily-work moment. Bare invocation shows a 5-station menu (starting / working / filing findings / reporting / wrapping up); with an argument (start | work | file | report | wrap, synonyms accepted) it jumps straight to the station and hands off to the right skill: ado-backlog:my-work, the situational toolbox (grill-then-plan, debug-mantra, study-design-verify, naming-audit, fit-gap-analysis, problem-description, ticket-trace, scrutinize, review-pr, dual-verifier, drive-to-legacy, crm-archaeology, generating-test-cases), findings-to-ado-backlog or ado-create-work-items (github-backlog twins on request), management-talk, or invoice-generator.
Writes and edits short product and marketing copy, including landing pages, CTAs, onboarding strings, product descriptions, email subjects, UI state copy (errors, empty states, button labels), brand voice charts, and AI-ism cleanup. Use when asked to "write copy", "fix the copy", "make this shorter", "improve the CTA", "write the error message", "empty state copy", "name this button", "rewrite from first principles", "remove AI-isms", "clean up AI writing", "flag AI patterns", "define our tone of voice", or "write a voice guide". For blog posts use the external ghostwriter skill with platform blog; for slide copy use presentation-creator; for docs use docs-writing; for in-session assistant talk use eli5; for product behavior decisions use product-design.
Write marketing copy for landing pages, product descriptions, CTAs, emails, and app UI text.
How to work in a Writing Pipeline project (the `writing-pipeline` starter pack): a three-stage drafting flow, ideas → drafts → published. Read when the project has these folders, OR when asked to capture an idea, start or revise a draft, review a piece before it ships, or publish it. Carries the stage flow and review behaviors so that guidance does not live inside template bodies or folder descriptions. Complements the platform `open-knowledge` skill; does not replace it.
Advanced YouTube metadata optimizer. Generates Browse-primary and Search-primary title variants, entity-rich descriptions, tags, chapters with Key Moments schema, 15-second hook script, translated metadata for top markets, pinned comment, and end-screen plan. Use when user says "optimize my video", "rewrite my title", "write YouTube description", "generate tags", or provides raw video info asking for metadata.
Rewrite titles, headlines, repository descriptions, homepage hero lines, LinkedIn headlines, article titles, and section openers so they lead with credible authority before explaining the content.
Uploads a drafted User Story or Technical Story to Azure DevOps as a work item, populating Title, Description, Acceptance Criteria, Story Points, Priority, and Tags, and optionally linking a parent feature. Also supports creating a full Epic → Feature → Story hierarchy from a structured backlog, and updating existing work items in bulk. Use when the user says "push to ADO", "create in ADO", "upload to ADO", "add to ADO", "update ADO items", or similar after stories have been drafted in conversation.
Build text-to-speech applications using Qwen3-TTS, a powerful speech generation system supporting voice clone, voice design, and custom voice synthesis. Use when creating TTS applications, generating speech from text, cloning voices from audio samples, designing new voices via natural language descriptions, or fine-tuning TTS models. Supports 10 languages (Chinese, English, Japanese, Korean, German, French, Russian, Portuguese, Spanish, Italian).