Latest Agent Skills

Build production-ready LLM applications, advanced RAG systems, and intelligent agents with vector search, multimodal AI, and enterprise integrations.

Integrates multi-dimensional health data to detect anomalies, predict risks (hypertension, diabetes, cardiovascular), and generate personalized recommendations and interactive HTML reports.

Coordinates a squad of specialized agents for software development, routing tasks and managing context across requirements, planning, architecture, implementation, review, testing, and deployment phases.

Orchestrates Gemini 3 Flash to audit Linux Kernel patches for memory safety, detect cost drifts in Terraform plans, and generate hardened Kubernetes security contexts.

Generate, iterate, and scale paid ad creative for Google Ads, Meta, LinkedIn, TikTok, and Twitter/X, using performance data to optimize headlines, descriptions, and primary text.

Provides a comprehensive reference and usage guide for the rclone command-line cloud storage manager, covering configuration, core commands, filtering, and safety practices.

Optimize post-signup onboarding, user activation, first-run experience, and time-to-value with structured audits, flow design, and measurement plans.

Reviews MDPR Markdown-to-presentation workflows with semantic hints, visual checks, and deterministic renderer boundaries.

Compares two code versions for semantic equivalence by tracing both side-by-side, identifying divergences, and producing a structured verdict.

Generates Jest unit and integration tests in JavaScript or TypeScript, covering mocking, snapshots, async testing, and React component testing.

Keeps a Mac awake during long builds, downloads, or automation runs using the built-in caffeinate command.

Drafts structured goal-loop prompts for long-running agent work with verifiable stop conditions, validation commands, and documentation requirements.

Scrapes public web data from LinkedIn, Twitter, GitHub, YouTube, and other sources, sends and reads email, performs deep research, generates images, and searches the web through the DeepAPI service.

Control cmux workspaces, panes, surfaces, and agent sessions safely from macOS terminal workflows.

Simulates a board of legendary marketing advisors to provide multiple expert perspectives on marketing questions, highlighting disagreements and synthesizing recommendations.

Model and solve LP, MILP, and QP problems using NVIDIA cuOpt's GPU-accelerated solver via Python, C/C++, or CLI interfaces.

Creates SageMaker endpoints (real-time or async) with autoscaling, CloudWatch alarms, and tagging enabled by default, then smoke-tests them before declaring success.

Plans and coordinates the deployment of a model to Amazon SageMaker AI, selecting the appropriate pathway (real-time, serverless, async, batch, or Bedrock CMI) based on model type, traffic, latency, and cost constraints.

Selects the correct SageMaker serving container image URI for HuggingFace model deployments, prioritizing HuggingFace-curated Deep Learning Containers over generic alternatives.

Discovers, validates, or creates a SageMaker execution role before deploying or training, preventing IAM-related deployment failures.

Reads the local AWS configuration to determine the active profile, region, account ID, and caller identity before any AWS task, avoiding guesswork and preventing common deployment errors.

Sets up an isolated Python environment for SageMaker and AWS work, pinning Python version and installing current boto3/awscli to avoid dependency conflicts and stale SDKs.

Designs and implements a Google Cloud solution for live, bidirectional multimodal streaming workloads with AI agents, covering requirements discovery, architecture design, and deployment planning.

Provides guidance on enabling and optimizing GKE Cluster Autoscaler, including Node Auto Provisioning, troubleshooting scale-up/down issues, and best practices for capacity management.

Compress natural language memory files (CLAUDE.md, todos, preferences) into a concise format to save input tokens while preserving technical substance, code, URLs, and structure.

Produces ultra-compressed, one-line code review comments that cut noise while preserving actionable signal. Each comment includes location, problem, and fix.

Generates ultra-compressed commit messages in Conventional Commits format with a subject line ≤50 characters and a body only when the 'why' isn't obvious.

Shows real token usage and estimated savings for the current Claude Code session by reading the session log.

Displays a quick-reference card for all caveman modes, skills, and commands. One-shot display, not a persistent mode.

Decides when to delegate to compressed-output subagents (investigator, builder, reviewer) instead of using vanilla agents, saving main-context tokens across long sessions.

Compresses AI agent output tokens by up to 65% using a terse, article-free communication style while preserving full technical accuracy. Supports multiple intensity levels and auto-disables for security warnings or ambiguous instructions.

Generates premium, app-native mobile screen images and flow images for iOS, Android, and cross-platform products, with strong emphasis on readability, consistency, and art-directed design.

Generates premium, conversion-aware website design reference images with one separate horizontal image per section, enforcing composition variety, narrative concept spine, and consistent palettes for landing pages and marketing sites.

Generates premium website design images, then deeply analyzes them to implement faithful frontend code for hero sections, landing pages, and marketing sites.

Generates ultra-minimalist, editorial-style web interfaces with warm monochrome palettes, typographic contrast, bento-grid layouts, and muted pastel accents, rejecting generic SaaS design trends.

Build web interfaces that fuse Swiss typographic design with military terminal aesthetics, using rigid grids, extreme type scale contrast, utilitarian color, and analog degradation effects for data-heavy dashboards, portfolios, or editorial sites.

Generates premium React/Next.js frontend code with deterministic typography, calibrated color palettes, and motion physics, enforcing anti-cliché design rules and performance guardrails.

Audits and upgrades existing websites and apps to premium quality by identifying generic AI patterns and applying high-end design standards without breaking functionality.

Generates award-level UI with GSAP motion, bento grids, and Python-driven randomization for layout variety.

Generates DESIGN.md files for Google Stitch that encode premium, anti-generic UI standards including typography, color, layout, motion, and anti-patterns.

Enforces complete, unabridged code generation by banning placeholder patterns and handling token-limit splits cleanly.

Provides design instructions for creating premium, agency-level visual interfaces with specific typography, spacing, shadows, card structures, and animations.

Generates premium brand-kit images including logo systems, identity decks, and visual-world presentations. Creates intentional, minimal, and strategic brand boards with grid-based layouts, symbolic logo concepts, and application mockups.

Finds exploitable vulnerabilities in codebases through a structured multi-phase audit covering recon, hunting, validation, and reporting.

Prevents AI agents from calling methods, imports, or variables that do not provably exist in the user's installed version, reducing hallucinated API usage.

Enforces a strict efficiency ladder—YAGNI, reuse, stdlib, native platform, existing deps—before writing any new code, preventing over-engineering and unnecessary dependencies.

Install and manage Python packages using uv pip when imports fail or dependencies are missing.

Create new skills, modify and improve existing skills, and measure skill performance through iterative evaluation and benchmarking.

Search the web and fetch content from URLs, returning LLM-friendly markdown.

Read, write, edit, list, copy, move, and search files directly using builtin tools, with fallback to Python or bash for complex operations.

Open, create, read, analyze, edit, or validate Excel/spreadsheet files (.xlsx, .xlsm, .csv, .tsv) with full formula support and professional formatting.

Create, read, edit, and convert .pptx presentations using python-pptx and PptxGenJS, with design guidance for professional slide decks.

Create, read, edit, and manipulate Word documents (.docx) with formatting, tables, images, headers, footers, tracked changes, and comments.

Read, extract, merge, split, rotate, encrypt, and create PDF files. Supports markdown-to-PDF conversion, table extraction, form filling, and OCR for scanned documents.

Add visual effects to Three.js scenes using EffectComposer: bloom, depth of field, anti-aliasing, ambient occlusion, and custom screen-space shaders.

Set up 3D scenes with Three.js, including scene, camera, renderer, Object3D hierarchy, coordinate systems, and math utilities.

Handle user input in Three.js scenes with raycasting, mouse/touch interaction, and camera controls including Orbit, Fly, FirstPerson, PointerLock, Trackball, and Map controls.

Create and configure Three.js materials including PBR, basic, phong, shader, and physical materials for 3D rendering.

Create and control Three.js animations including keyframe, skeletal, morph target, and animation blending.

Load, configure, and apply textures in Three.js, including UV mapping, environment maps, HDR textures, and render targets.