Loading...
Loading...
Found 1,389 Skills
Ultra-lightweight channel for refactor processes - used when changes are clearly too small to go through the full scan → design → apply three-stage workflow. AI directly identifies 1-3 low-risk optimization points, confirms with the user once, modifies in-place using classic methods, and validates itself by running tests. No scan checklist, no design documentation, no multi-step human verification required. Trigger scenarios: User says "quick refactor", "small refactor", "simply optimize XX function", "modify directly", "skip the extra steps", and the scope of changes is clearly localized to a single function / single component with test coverage for self-validation.
Enter this sub-process when conducting code optimization — handle tasks where 'behavior remains unchanged, structure changes' (structure / performance / readability). Shift single-module internal optimization from 'AI random refactoring' to 'first scan to generate a checklist, confirm each item with the user, execute step-by-step according to the method library, and require manual approval for each step'. Trigger scenarios: Users mention phrases like 'optimize it / refactor / rewrite / split it / poor performance / code is too long' without any accompanying behavior changes. Do not handle new requirements (route to feature), bugs (route to issue), or cross-module architecture restructuring (route to architecture + decisions).
Generate complete, accessible color palettes from a single brand hex. Creates 11-shade scale (50-950), semantic tokens (background, foreground, card, muted), and dark mode variants. Includes WCAG contrast checking for text accessibility. Use when: setting up design system, creating Tailwind theme, building brand colors from single hex, converting designs to code, checking color accessibility.
Build and deploy Streamlit apps natively in Snowflake. Covers snowflake.yml scaffolding, Snowpark sessions, multi-page structure, Marketplace publishing as Native Apps, and caller's rights connections (v1.53.0+). Use when building data apps on Snowflake, deploying SiS, fixing package channel errors, authentication issues, cache key bugs, or path resolution errors.
Multi-provider email sending for Cloudflare Workers and Node.js applications. Build transactional email systems with Resend (React Email support), SendGrid (enterprise scale), Mailgun (developer webhooks), or SMTP2Go (reliable relay). Includes template patterns, webhook verification, attachment handling, and error recovery. Use when sending emails via API, handling bounces/complaints, or migrating between providers.
Comprehensive code review skill for TypeScript, JavaScript, Python, Swift, Kotlin, Go. Includes automated code analysis, best practice checking, security scanning, and review checklist generation. Use when reviewing pull requests, providing code feedback, identifying issues, or ensuring code quality standards.
Foundational plotting library. Create line plots, scatter, bar, histograms, heatmaps, 3D, subplots, export PNG/PDF/SVG, for scientific visualization and publication figures.
Comprehensive guide for architecting Flutter applications following MVVM pattern and best practices with feature-first project organization. Use when working with Flutter projects to structure code properly, implement clean architecture layers (UI, Data, Domain), apply recommended design patterns, and organize projects using feature-first approach for scalable, maintainable apps.
Deploy Python applications to Google App Engine Standard/Flexible. Covers app.yaml configuration, Cloud SQL socket connections, Cloud Storage for static files, scaling settings, and environment variables. Use when: deploying to App Engine, configuring app.yaml, connecting Cloud SQL, setting up static file serving, or troubleshooting 502 errors, cold starts, or memory limits.
Compress large language models using knowledge distillation from teacher to student models. Use when deploying smaller models with retained performance, transferring GPT-4 capabilities to open-source models, or reducing inference costs. Covers temperature scaling, soft targets, reverse KLD, logit distillation, and MiniLLM training strategies.
Train Mixture of Experts (MoE) models using DeepSpeed or HuggingFace. Use when training large-scale models with limited compute (5× cost reduction vs dense models), implementing sparse architectures like Mixtral 8x7B or DeepSeek-V3, or scaling model capacity without proportional compute increase. Covers MoE architectures, routing mechanisms, load balancing, expert parallelism, and inference optimization.
GPU-accelerated data curation for LLM training. Supports text/image/video/audio. Features fuzzy deduplication (16× faster), quality filtering (30+ heuristics), semantic deduplication, PII redaction, NSFW detection. Scales across GPUs with RAPIDS. Use for preparing high-quality training datasets, cleaning web data, or deduplicating large corpora.