Skills worth trusting.
Agent skills from Anthropic, OpenAI, skills.sh, ClawHub, SkillsMP and other reputable sources. Savant scans each one with NVIDIA SkillSpector and evaluates it live, with an LLM drafting and running test cases and Jev validating and scoring them. Workspaces import any skill into their own repositories through a reviewed pull request. Every listing on each hub is enumerated; packages are fetched, scanned and evaluated most-popular first as daily capacity allows.
ui-designer
Craft one-of-a-kind, ship-ready frontend interfaces with elevated design quality. Apply this skill whenever the user wants to build web components, pages, artifacts, posters, or applications (for example, websites, landing pages, dashboards, React components, HTML/CSS layouts, or any task that involves styling or beautifying a web UI). Produces inventive, refined code and UI design that sidesteps cookie-cutter AI aesthetics.
testing-agentforce
Write, run, and analyze structured test suites for Agentforce agents. TRIGGER when: user writes or modifies test spec YAML (AiEvaluationDefinition); runs sf agent test create, run, run-eval, or results commands; asks about test coverage strategy, metric selection, or custom evaluations; interprets test results or diagnoses test failures; asks about batch testing, regression suites, or CI/CD test integration. DO NOT TRIGGER when: user creates, modifies, previews, or debugs .agent files (use developing-agentforce); deploys or publishes agents; writes Agent Script code; uses sf agent preview for development iteration; analyzes production session traces (use observing-agentforce).
video-frames
Extract frames or short clips from videos using ffmpeg.
observing-agentforce
Analyze production Agentforce agent behavior using session traces and Data Cloud. TRIGGER when: user queries STDM session data or Data Cloud trace records; investigates production agent failures, regressions, or performance issues; asks about session traces, conversation logs, or agent metrics; wants to reproduce a reported production issue in preview; runs findSessions or trace analysis queries. DO NOT TRIGGER when: user creates, modifies, or debugs .agent files during development (use developing-agentforce); writes or runs test specs (use testing-agentforce); uses sf agent preview for local development iteration; deploys or publishes agents.
loop-design-check
Design a goal-oriented agent loop or review one for failure modes: spinning, Goodhart-gaming the verifier, or running a wrong answer to completion. Covers machine-decidable goals, loop types, plan/build/judge skeletons, and runaway prevention; mechanism wiring lives in autonomous-loops. Use when designing, writing, or checking an agent loop. 中文触发：写 loop、设计 loop、做一个 loop、检查 loop 对不对、loop 体检、loop 会不会跑飞、可判定目标、五个崩法、plan build judge。
xcode-build-orchestrator
Orchestrate Xcode build optimization by benchmarking first, running the specialist analysis skills, prioritizing findings, requesting explicit approval, delegating approved fixes to xcode-build-fixer, and re-benchmarking after changes. Use when a developer wants an end-to-end build optimization workflow, asks to speed up Xcode builds, wants a full build audit, or needs a recommend-first optimization pass covering compilation, project settings, and packages.
shadcn-ui
Install and configure shadcn/ui components for React projects. Guides component selection, installation order, dependency management, customisation with semantic tokens, and common UI recipes (forms, data tables, navigation, modals). Use after tailwind-theme-builder has set up the theme infrastructure, when adding components, building forms, creating data tables, or setting up navigation.
platform-encryption-configure
Configure Salesforce Shield Platform Encryption — generate deployable encryption settings and encrypted-field metadata, and answer key-model and lifecycle questions. TRIGGER when: user wants to turn on deterministic encryption, encrypt a field, set up Cache-Only Keys, External Key Management, or replay detection, or mentions Shield Platform Encryption, encryption at rest, deterministic vs probabilistic encryption, encryptionScheme, PlatformEncryptionSettings, EncryptionKeySettings, BYOK, BYOKMS, tenant secrets, key rotation, or .settings-meta.xml / .field-meta.xml for encryption — even when they don't say 'Shield'. SKIP when: user needs a generic custom field with no encryption (use platform-custom-field-generate), needs the raw Metadata API type reference (use platform-metadata-api-context-get), or asks about Classic Encryption (encrypted text fields), which is a different feature. Use this skill for any Platform Encryption configuration, field-encryption, or key-model question.
bmad-toolsmith
Smithy the Toolsmith builds, converts, reviews, packages, validates and migrates BMad skills, agents and modules. Use when the user wants a skill, agent, persona, command or module made, changed, reviewed or packaged; a prompt or rule from another tool turned into a skill; a skill mined from session logs; an old module migrated; or asks for Smithy. Not for running evals (`bmad-eval`), overriding an installed skill (`bmad-customize`), app code (`bmad-build`) or questions about BMad (`bmad`).
using-ui-bundle-salesforce-data
MUST activate when the project contains a uiBundles/*/src/ directory and the task involves ANY Salesforce record operation — reading, creating, updating, or deleting. Use this skill when building forms that submit to Salesforce, pages that display Salesforce records, or any code that touches Salesforce objects or custom objects. Activate when files under uiBundles/*/src/ import from @salesforce/sdk-data, or when *.graphql files or codegen.yml exist. This skill owns all Salesforce data access patterns in UI bundles. Does not apply to authentication/OAuth setup, schema changes, Bulk/Tooling/Metadata API, or declarative automation.
image-taste-frontend
Elite website image-to-code skill for Codex. For visually important web tasks, it must first generate the design image(s) itself, deeply analyze them, then implement the website to match them as closely as possible. In Codex, it must prefer large, readable, section-specific images instead of tiny compressed boards, generate fresh standalone images for sections or detail views instead of cropping old ones, avoid lazy under-generation, avoid cards-inside-cards-inside-cards UI, and keep the hero clean, spacious, readable, and visible on a small laptop.
bmod-toolsmith
Required bmod metadata. Never invoke this skill.
react-router-framework-mode
Build full-stack React applications using React Router's framework mode. Use when configuring routes, working with loaders and actions, handling forms, handling navigation, pending/optimistic UI, error boundaries, or working with react-router.config.ts or other react router conventions.
typescript-discriminated-unions
Apply, review, and explain discriminated unions in TypeScript. Use automatically for tasks involving mutually exclusive variants, invalid state prevention, exhaustiveness checking, variant-specific object properties, application state, function arguments, or React props.
migrating-motoko-actors
Motoko actor migration and schema evolution with the enhanced migration chain (migrations/). Load when upgrading canisters or changing actor state shape.
troubleshooting-motoko-migrations
Deep reference for when the migration chain misbehaves — upgrade compatibility errors you cannot explain, migration files you cannot write, a project converted from legacy persistence, or a request to remove the chain. Load only when the rules in migrating-motoko-actors do not explain what you are seeing.
dry-refactoring
Guided workflow to eliminate copy-paste duplication detected by jscpd. Refactor exact, renamed and near-miss clones using extract function, parameterize, module, constant, or base class strategies, starting from the hotspots the summary ranks.
ai-behavior-trees-utility-ai
Build a production behavior-tree runtime (Blackboard, action/condition leaves, sequence/selector/parallel composites, decorators) and a Utility AI system (response curves — linear, exponential, sigmoid, quadratic — considerations, and action evaluators), plus hybrid BT-drives-Utility agents. Use when implementing a reusable behavior-tree or utility-based decision system, or tuning enemy/NPC decisions beyond a simple FSM, or when the user mentions behavior tree, blackboard, decorator, selector, sequence, tick status, utility AI, response/scoring curve, or consideration. For choosing between FSM/BT/steering or for pathfinding, use game-ai; for Unreal's BehaviorTree/Blackboard assets, use unreal-behavior-trees.
project-references
Look up conventions, patterns, and concrete implementations from your own GitHub repositories checked out locally under ~/projects/referenzen/. Use this skill whenever there is uncertainty about how something is done in your codebase family — e.g. Helm chart structure, Kubernetes manifests, framework configuration patterns, Docker Compose conventions, CI/CD pipeline setup, or any other recurring architectural decision. Invoke it proactively before guessing at a convention; always cite the source project and path when a pattern is adopted. Also use when the user asks to check out, update, or search reference repositories, or to generate an overview/report of GitHub Actions triggers (push, pull_request, schedule/cron) across repositories.
implementing-ui-bundle-file-upload
MUST activate when the project contains a uiBundles/*/src/ directory and the task involves uploading, attaching, or dropping files. Use this skill when adding file upload functionality to a UI bundle app. Provides progress tracking and Salesforce ContentVersion integration. This feature provides programmatic APIs ONLY — build custom UI using the upload() API. ALWAYS use this instead of building file upload from scratch with FormData or XHR.
typescript-variables
Apply, review, and explain TypeScript variable conventions. Use automatically for tasks involving variable declarations, const assertions, enum alternatives, literal-union state modelling, boolean flags, or choosing between null and undefined.
gke-ai-troubleshooting-tpu-metrics-monitoring
Monitors and troubleshoots GKE TPU workloads, nodes, and node pools using GKE system metrics and PromQL. Use when monitoring TensorCore duty cycle, TPU memory, node readiness, multi-host TPU node pool availability, host maintenance or preemption interruptions, and calculating MTTR or MTBI metrics for GKE TPUs. Don't use for general non-TPU GKE workload monitoring or non-metric TPU debugging.
directory-build-organization
Guide for organizing MSBuild infrastructure with Directory.Build.props, Directory.Build.targets, Directory.Packages.props, and Directory.Build.rsp. USE FOR: structuring multi-project repos, centralizing build settings, implementing NuGet Central Package Management (CPM) with ManagePackageVersionsCentrally, consolidating duplicated properties across .csproj files, setting up multi-level Directory.Build hierarchy with GetPathOfFileAbove, understanding evaluation order (Directory.Build.props → SDK .props → .csproj → SDK .targets → Directory.Build.targets). Critical pitfall: $(TargetFramework) conditions in .props silently fail for single-targeting projects — must use .targets. DO NOT USE FOR: non-MSBuild build systems, migrating legacy projects to SDK-style (use msbuild-modernization), single-project solutions with no shared settings.
open-code-review-delegate
Delegation mode for open-code-review (OCR). Instead of OCR calling an LLM endpoint, this skill instructs the host agent to perform the code review itself, using OCR only for deterministic engineering: file selection and rule resolution. Use when the host agent should drive the review with its own LLM capabilities.
msbuild-modernization
Guide for modernizing and migrating MSBuild project files to SDK-style format. USE FOR: converting legacy .csproj/.vbproj with verbose XML to SDK-style, migrating packages.config to PackageReference, removing Properties/AssemblyInfo.cs in favor of auto-generation, eliminating explicit <Compile Include> lists via implicit globbing, consolidating shared settings into Directory.Build.props. Indicators of legacy projects: ToolsVersion attribute, <Import Project="$(MSBuildToolsPath)">, .csproj files > 50 lines for simple projects. DO NOT USE FOR: projects already in SDK-style format, non-.NET build systems (npm, Maven, CMake), .NET Framework projects that cannot move to SDK-style.
coverage-analysis
Activation requires either supplied .NET coverage reports/percentages/line, branch, or condition metrics, or an explicit request to collect .NET coverage for analysis. USE FOR: interpreting that evidence, including Cobertura data, partial conditions, plateaus, target arithmetic, project-wide coverage-backed CRAP, and coverage-backed refactoring safety. Analyze supplied reports directly without rerunning tests or installing tools. DO NOT USE FOR: requests with neither coverage evidence nor explicit coverage-collection intent, including hypothetical change-survival questions (use test-gap-analysis); CRAP or refactoring safety for one named target (use crap-score); or requests owned by test-tagging, find-untested-sources, test-anti-patterns, run-tests, or code-testing-agent.
gjkim-instruction
Create and maintain gjkim_instruction.md, the root document for a loop-engineering effort. The document holds only the minimum requirements and confirmed decisions, and deliberately leaves everything else open so that later iterations are not locked into early guesses. Use whenever the user mentions gjkim_instruction.md, a root document or root doc for a loop, loop engineering, starting a long-running agent loop on a goal, or asks to record a requirement or a confirmed decision for such an effort. Also use at the start of a loop iteration to check what is already decided versus still open.
dotnet-aot-compat
Make .NET projects compatible with Native AOT and trimming by systematically resolving IL trim/AOT analyzer warnings. USE FOR: making projects AOT-compatible, fixing trimming warnings, resolving IL warnings (IL2026, IL2070, IL2067, IL2072, IL3050), adding DynamicallyAccessedMembers annotations, enabling IsAotCompatible. DO NOT USE FOR: publishing native AOT binaries, optimizing binary size, replacing reflection-heavy libraries with alternatives. INVOKES: no tools — pure knowledge skill.
dart-use-path-package
Cross-platform file and directory path manipulation, segment splitting, extension extraction, and context conversion using `package:path` and `package:file`. Use when writing, inspecting, joining, splitting, or refactoring file paths, directory names, or extensions, or replacing raw string path operations (`.split('/')`, `'$dir/$file'`, `.endsWith('.ext')`, `.replaceAll('\\', '/')`). Don't use for HTTP network URI routing, database query strings, or non-path string processing.
correct
Find the mistakes agents keep repeating in this repo and make each one impossible. Try architecture first, then types, then a lint whose error names the fix, then a test, and write docs last. Prove each check fails on a real past mistake. Repeat this each time the operator corrects you. Use for /correct.
kie-models
在 KIE 平台(api.kie.ai)上发现、查价并调用 AI 模型。当用户提到 KIE 或 kie.ai,问"X 用哪个 KIE 模型"、"列出 KIE 模型"、"在 KIE 上跑模型",或需要某个模型的 schema、价格、成功率时使用。
design-cleanup
Sweep the current branch for leftovers of an abandoned design after a significant mid-work decision change, and rewrite the affected code, comments, tests, and docs so the result reads as if it had been designed this way from the beginning. Use when the user says "design cleanup", "clean up the old approach", "erase the journey", says a design decision changed and asks to clean up after it, or after a pivot in approach mid-branch. This is a rewrite pass over the code touched by the decision only — not a general refactoring or bug-hunting pass.
gravitas
Use for software implementation, debugging, refactoring, review, migration, and other repository-changing engineering work where correctness must be verified. Enforces scoped reconnaissance, adaptive planning, evidence-backed verification, requirement tracking, failure recovery, and budget-aware delegation. Model-agnostic reliability runtime for coding agents -- measured, not claimed.
ruff
Guide for using ruff, the extremely fast Python linter and formatter. Use this when linting, formatting, or fixing Python code.
bro
Restate the last message in plain human language, with no jargon.
principle-explain-the-number
Apply before you trust, report, or act on a number you measured: a speedup, a regression, a throughput, a latency, or an eval result. Find what limits it, and rule out that it measured something other than the work you think.
qoder-delegate
Delegate a coding task to the Qoder CLI (`qodercli`) as a background implementer, then review its diff and land it yourself. Use this whenever the user asks to have Qoder implement, fix, refactor, or run a queue of coding tasks while the orchestrator remains the reviewer. DO NOT USE for tasks small enough to do inline, or when the user wants code written directly without delegation.
k6-perf-test-website
Use when the user wants to performance-test, load-test, or stress-test a public website end-to-end with k6. Produces a hybrid (protocol + browser) test suite, SLO-backed thresholds, a load-generator monitor sidecar, and a Grafana-side investigation playbook for backends the user owns. Triggers on "perf test my site", "performance test my site", "load test this URL", "stress test my web app", "I want to load test [URL]", "set up k6 against my website", "write a k6 suite for [site]", "see if my site handles N concurrent users", or "how does my site perform under traffic". Use this skill whenever the user mentions k6, load testing, stress testing, performance testing, or wants to validate a website under traffic — even if they don't explicitly use the word "test" or ask for the specific outputs this skill produces.
benchmark-checklist
Vet a perf measurement (limiter, tuning, limits, errors, repeatability, relevance, and whether the work happened) before you report or act on it. Use when you run a benchmark or report a speedup or regression you measured.
steve-jobs-design-review
Review designs, products, and features with Steve Jobs' standards: ruthless simplicity, focus, and end-to-end excellence. Use when the user mentions "Steve Jobs review", "design review", "product review", "what would Steve do", "insanely great", "this feels too complicated", "too many features", "product taste", "saying no", or "is this good enough to ship". Also trigger when critiquing a UI, feature, or roadmap for focus and simplicity, cutting scope to the essential, or pressure-testing the whole experience from first run to daily use. Covers the simplicity audit, the no list, design-is-how-it-works, end-to-end ownership, demo culture, and a Jobs-style review protocol with binary verdicts. For visual design fundamentals, see refactoring-ui. For usability audits, see ux-heuristics. For detail polish, see microinteractions.
cc-best-practices
Guidance on how to use Claude Code effectively — covering context management, verification strategies, the explore-plan-implement workflow, prompting techniques, session management, parallel sessions, and common failure patterns. Use this skill whenever the user asks how to get the most out of Claude Code, how to write better prompts, how to manage context, when to use plan mode, how to automate tasks, or when they describe a frustrating pattern like Claude repeating mistakes or losing track of instructions.
netlify-config
Configure Netlify builds and routing via netlify.toml, _redirects, and _headers. Use when setting a build command or publish directory, adding redirects or rewrites or proxies, adding an SPA fallback rewrite, setting custom response headers or basic auth, managing environment variables and secrets, scoping vars per deploy context, marking a var as secret, disabling secret scanning, configuring functions bundling, ignoring builds, or wiring up a monorepo or JavaScript SPA on Netlify.
neon-realtime
Neon Realtime makes any SQL query live. Build realtime, reactive apps and agents on Lakebase Postgres with live SQL queries and reactivity primitives. Use when app UI or agent state must update live from database rows (dashboards, feeds, collaboration, multi-user or multi-device sync), for optimistic mutations, or to replace polling or other realtime sync systems. Triggers include "live query", "subscribe", "sync engine", "useLiveQuery", "TanStack DB", "optimistic updates", "@neon/realtime". For push that doesn't come from database rows (presence, typing, LLM streams), use neon-functions.
audio-setup-mixers
Routes Audio Sources into existing Audio Mixer Groups, classifying each source by what it plays. Use when the user asks about cleaning up mixer assignments, routing audio through a mixer, or which group a sound belongs in.
cx-observability-setup
Use this skill when the user asks to "set up monitoring", "configure observability", "onboard new service", "create saved view", "set up notifications", "configure webhook", "set up Slack integration", "outgoing webhook", "automation action", "webhook for alerts", "create view", "saved view", "view folder", "organize dashboards", "install integration", "configure extension", "contextual data", "connect external service", "create notification connector", "set up email alerts", "configure PagerDuty", "notification routing", "deploy extension", "test webhook", "notification preset", "test notification", "webhook actions", or wants to set up, configure, or manage the observability stack for a service or team.
threejs-gameplay-systems
Build and iterate playable Three.js game systems: starter scaffold, architecture, design briefs, core loops, level and encounter design, entities, input, camera, collision and physics, scoring, objectives, and game feel. Use for first playable slices, new Vite/TypeScript/Three.js setups, level/arena/track/wave/hole/puzzle design, combat encounters, difficulty tuning, and juice.
working-with-legacy-code
Safely change and test untested codebases using Feathers' "Working Effectively with Legacy Code". Use when the user mentions "legacy code", "no tests", "untested codebase", "how do I test this", "seams", "characterization tests", "golden master", "sprout method", "afraid to change this code", "monster method", "dependency breaking", or "inherited a messy codebase". Also trigger when changing code without tests safely, getting a class under test when constructors, statics, or singletons block it, adding features to tangled modules, or planning incremental test coverage for an old codebase. Covers the legacy-code change algorithm, seams, characterization tests, sprout/wrap, and dependency-breaking techniques. For refactoring code that already has tests, see refactoring-patterns. For day-to-day code quality, see clean-code.
vibe-delegate
Delegate a coding task to the Mistral Vibe CLI (`vibe`) as a background implementer, then review its diff and land it yourself. Use this whenever the user wants to hand implementation work to Vibe — phrasings like "have Vibe implement X", "delegate this to Vibe", "run it through Mistral Vibe", "use vibe to implement/fix/refactor" — or wants to run a queue of coding tasks through Vibe while staying the reviewer. DO NOT USE for tasks small enough to do inline, or when the user wants the code written directly without delegating.