Top Agent Skills

Detect memory errors like buffer overflows and use-after-free bugs in C/C++ code during fuzzing and testing using AddressSanitizer.

Enables ultra-granular, line-by-line code analysis to build deep architectural context before vulnerability or bug finding.

Audits GitHub Actions workflows for security vulnerabilities in AI agent integrations, detecting attack vectors where attacker-controlled input reaches AI agents in CI/CD pipelines.

Generate songs, instrumental tracks, and covers using the MiniMax Music API with basic or advanced control modes.

Discover, bet on, track, and settle Hunch prediction markets in natural language with on-chain settlement on Base via x402.

Play an onchain capture-the-flag game on Base by yoinking the flag from the current holder, checking game stats and leaderboards, and competing for the trophy using Bankr for transaction execution.

Perform efficient pandas DataFrame operations for data analysis, manipulation, and transformation with production-grade patterns.

Writes, reviews, and debugs idiomatic Rust code with memory safety and zero-cost abstractions, covering ownership patterns, lifetimes, trait hierarchies, async/await with tokio, and error handling with Result/Option.

Builds security-focused full-stack web applications by implementing integrated frontend and backend components with layered security at every level, covering the complete stack from database to UI.

Designs distributed system architectures, decomposes monoliths into bounded-context services, recommends communication patterns, and produces service boundary diagrams and resilience strategies.

Design tool descriptions, schemas, and response formats that agents can reliably route on, including MCP server design, tool-set consolidation, and naming conventions.

Store and query embeddings with metadata filtering, vector search, and full-text search using an open-source database that scales from notebooks to production.

Quantize large language models to 4-bit using activation-aware weight quantization, achieving ~3x speedup with minimal accuracy loss for deployment on limited GPU memory.

Fine-tune and serve Physical Intelligence OpenPI models (pi0, pi0-fast, pi0.5) using JAX or PyTorch backends for robot policy inference across ALOHA, DROID, and LIBERO environments.

Train large-scale MoE models with FP8/INT4 low-precision RL, speculative decoding, and train-inference alignment using the miles framework.

Generate music and sound effects from text descriptions using Meta's AudioCraft library, with support for melody conditioning, stereo output, and style transfer.

Run ML training and batch jobs across multiple clouds with automatic cost optimization, spot instance recovery, and unified orchestration.

Train and use SentencePiece tokenizers for multilingual NLP, supporting BPE and Unigram algorithms with raw Unicode text.

Pretrains large language models from scratch using PyTorch-native distributed training with 4D parallelism (FSDP2, TP, PP, CP) and Float8 support on H100 GPUs.

Fast tokenization for NLP using Rust-based tokenizers supporting BPE, WordPiece, and Unigram algorithms, with training, alignment tracking, and padding/truncation.

Manage multiple n8n instances over MCP by discovering, switching, and verifying the target instance before reads and writes.

Detect when backtest iteration has stalled and generate structurally different strategy pivot proposals to break out of local optima.

Clusters raw detection tickets into reusable edge concepts with thesis, invalidation signals, and strategy playbooks before strategy design.

Fetch upcoming economic events and data releases using the FMP API, including central bank decisions, employment reports, inflation data, GDP releases, and other market-moving indicators for specified date ranges.

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

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

Diagnose and validate cron expressions before they ship, catching impossible dates, OR-semantics, midnight spikes, uneven step drift, and leap-year February 29.

SKILL — Design System

Use for writing, reviewing, debugging, testing, or diagnosing Bayesian regression and hierarchical models built with Bambi formulas, Model, Family/Likelihood/Link, Prior, fit, prior predictive, and predict. Trigger on common versus group-specific terms, categorical coding, family/link choice, automatic prior scaling, missing rows, PyMC backend settings, and InferenceData predictions. Do not use for hand-built PyMC graphs, NumPyro programs, ArviZ-only analysis of existing draws, frequentist statsmodels formulas, or generic pandas work.

`analysis-agent`: use when a request needs normal, failure, edge, abuse, recovery, or operational scenarios; skip when no scenario-decomposition decision exists.

Socratic questioning protocol + user communication.

Data pipeline specialist for ETL design, data quality, CDC patterns, and batch/stream processingUse when "data pipeline, etl, cdc, data quality, batch processing, stream processing, data transformation, data warehouse, data lake, data validation, data-engineering, etl, cdc, batch, streaming, data-quality, dbt, airflow, dagster, data-pipeline, ml-memory" mentioned.

Regulatory compliance and audit agent. Maps business regulatory requirements (SOC2/PCI-DSS/HIPAA/ISO 27001), checks control implementations, designs audit trails, and implements Policy as Code. Use when compliance auditing is needed.

Run and control a user's app on a remote iOS/Android simulator hosted on EAS cloud, enabling interaction from environments without local simulators.

Execute AEM Edge Delivery Services admin operations: manage content, cache, code, indexes, sitemaps, snapshots, logs, users, jobs, sites, config, secrets, API keys, tokens, profiles, and versioning. Also supports Document Authoring operations.

Generates a compact HTML KPI digest from Adobe Analytics, showing how key metrics changed over a period and what drove the movement.

Creates complete AEM components with dialog, HTL template, Sling Model, unit tests, and clientlibs. Supports extending Core Components and project components, and can translate Figma designs into pixel-perfect HTL, CSS, and JS.

Generate comprehensive admin documentation for AEM Edge Delivery Services project handover, covering Config Service setup, permissions, Admin API operations, and troubleshooting.

Import a single webpage from any URL into canonical EDS block format — structured HTML that authors edit in DA. Scrapes the page, analyzes structure, maps to existing blocks, and generates HTML for immediate local preview.

Hand off tasks to the SLICC browser agent or install a new skill into SLICC from a GitHub repository.

Guides migration of AEM Cloud Service schedulers from legacy patterns to OSGi-compliant Sling Scheduler or Sling Jobs, with classification, review checklists, and troubleshooting.

Orchestrates end-to-end lifecycle workflows for the Adobe Dispatcher Apache HTTP Server module in AEM 6.5 LTS, from design and implementation through validation, release readiness, and incident troubleshooting.

Build evaluation frameworks for agent systems with deterministic checks, regression suites, multi-dimensional rubrics, quality gates, production monitoring, and outcome measurement.

Model agent mental states using BDI (Beliefs, Desires, Intentions) ontology patterns, enabling cognitive reasoning, explainability, and semantic interoperability in multi-agent systems.

Makes n8n workflow failures loud, structured, and recoverable by wiring per-node error outputs, retry logic, error-trigger workflows, and proper HTTP response shapes.

Audits a design against Dieter Rams' ten principles, scores each with evidence, and produces a verdict with a handoff prompt for new, refined, or redesigned output.

Fine-tune Gemma models locally using QLoRA, Unsloth, or TRL for SFT, DPO, and reward modeling, with dataset preparation and conversion to GGUF or LiteRT-LM.

Generate a Swiss International style presentation deck with 16-column grid, one of four accent colors, and 22 locked layout templates. Outputs a single self-contained HTML file.

Analyze and replicate xAI's design language: near-black canvas, white pill outlines, Universal Sans display type, and Geist Mono technical labels.

Provides a comprehensive design system specification inspired by Tesla's website, covering visual theme, color palette, typography, and UI patterns.

Documents the design system for Cursor's marketing site, including color palette, typography, layout, and component tokens.

Documents the design system behind Ollama's marketing site, including color tokens, typography hierarchy, spacing, and component patterns for a minimal, documentation-first aesthetic.

Analyzes and documents the design system of Together AI's website, extracting colors, typography, and surface patterns for use in design projects.

Provides a production-ready UI motion system for React/Next.js with performance, accessibility, and usability guidelines.

Deploys and optimizes AI/ML inference workloads on GKE, using GPUs, TPUs, and model servers.

Creates GKE clusters with golden path Autopilot defaults, supporting Standard and GPU workloads. Guides through project, region, and networking inputs, then provisions and verifies cluster settings.

Guides authentication and authorization to Google Cloud services, covering human users, service identities, Application Default Credentials (ADC), and best practices for secure access.

Evaluates Google Cloud workloads against the Well-Architected Framework security pillar, identifies security requirements, and provides actionable recommendations for IAM, network security, data protection, and operational security.

Fetches IAM recommendations and security insights from Google Cloud for a specified organization, folder, or project, using MCP tools, gcloud CLI, or direct API calls.

Guides a developer through first steps on Google Cloud, including account setup, authentication, project creation, and billing linkage.