openai/plugins
38 skills946 installs186k starsGitHub
Install
npx skills add https://github.com/openai/pluginsSkills in this repo
1Plugin CreatorCreate and scaffold plugin directories for Codex with a required `.codex-plugin/plugin.json`, optional plugin folders/files, and baseline placeholders you can edit before publishing or testing. Use when Codex needs to create a new personal plugin, add optional plugin structure, or generate or update personal or repo-root `.agents/plugins/marketplace.json` entries for plugin ordering and availability metadata. 2. Open `<plugin-path>/.codex-plugin/plugin.json` and replace `[TODO: ...]` placeholders.170installs2Android Emulator QaThe android-emulator-qa skill documents a repeatable Android emulator testing workflow for mobile apps. It starts by listing devices with adb devices, selecting a serial, and optionally cold-booting an AVD when no emulator is connected. Build and install steps run Gradle install tasks such as ./gradlew :module:installBuildVariant, resolve the launch activity with adb shell cmd package resolve-activity, and start the app with am start. Visual verification captures screenshots via adb exec-out screencap. Interaction commands cover tap, swipe away from screen edges, text input, back keyevents, and uiautomator UI tree dumps. Coordinate picking always derives centers from UI tree bounds using the bundled ui_pick.py helper, re-dumping after scrollable searches before concluding a target is missing. A ui_tree_summarize.py helper produces compact skeletons before reading full XML. Logcat guidance clears buffers, streams per-process logs by pid, reads crash buffers, and saves dumps for later review. Package shortcuts filter installed apps and confirm activities before launch. Use for Android emulator QA, UI automation, and crash investigation.124installs3Frontend App BuilderThe frontend-app-builder skill creates polished frontend apps, dashboards, games, and creative websites by acting first as a senior designer and then as an implementation engineer. Hard rules require Image Gen concepts unless the user opts out, complete surface design before coding, section-specific references for multi-section sites, and agency-signoff fidelity between accepted concepts and browser screenshots verified with view_image. It builds a design system from tokens, typography, component families, spacing, icon inventory, and container rules before writing code, defaulting complex app UIs to React plus Vite unless the repo dictates otherwise. Prohibited defaults include hero eyebrow pills, generic card grids, creative liberties on layout or copy, and handoff before concept versus implementation comparison passes. Browser plugin verification is primary with Playwright Chromium as fallback when noted. The workflow coordinates image generation briefing, design approval in plan mode, faithful implementation, responsive fixes, and removal of temporary QA artifacts. Triggers include new frontend apps, dashboards, hero sections, redesigns, and visually driven UI from scratch.70installs4Ios Memgraph LeaksThe ios-memgraph-leaks skill helps prove iOS memory leaks from live simulator processes or existing memgraph files using bundled capture and summarize scripts. The workflow builds and drives the reproduction flow, captures a memgraph with capture_sim_memgraph.sh using simulator UDID and bundle identifier, summarizes leaks via summarize_memgraph_leaks.py, and inspects app-owned types with leaks traceTree or groupByType evidence. Fixes require the smallest root-cause patch retaining-edge removal rather than broad cleanup, then recapture on the same simulator with before-and-after leak counts and disappeared types. Agents must not claim fixes from smaller memgraphs alone without explaining ownership paths. SKILL_DIR resolves to the loaded plugin folder, not the target app repo. Pair with ios-debugger-agent when builds, UI driving, logs, or screenshots are also needed. Triggers include debugging leaked objects, retain cycles, or memory growth on iOS simulators.56installs5Ios App IntentsThe ios-app-intents skill designs a minimal App Intents layer that exposes high-value iOS actions to Shortcuts, Siri, Spotlight, widgets, and controls. It starts from one to three verbs people want outside the app such as compose, open, find, filter, continue, inspect, or start rather than mirroring full navigation trees. AppEntity types stay narrower than persistence models with EntityQuery only where disambiguation helps. Each action chooses inline completion versus openAppWhenRun handoff into a predictable main-scene routing path instead of scattered side effects. AppShortcutsProvider entries use direct task-oriented phrases and symbols aligned across widgets and controls sharing parameters. The workflow validates compile success, runtime routing, and summarizes exposed actions, backing entities, and invocation handling. Strong defaults favor a dedicated intents target, thin intent types with domain logic in services, small display-friendly entities, AppEnum for fixed choices, and treating App Intents as system integration infrastructure. Anti-patterns include screen-per-intent taxonomies, full model graphs as entities, and vague shortcut phrases.55installs6Ios Ettrace PerformanceThe ios-ettrace-performance skill captures focused, symbolicated ETTrace profiles from iOS Simulator apps. The workflow picks one user-visible flow, builds the exact simulator binary, temporarily links ETTrace.xcframework into the app target, collects UUID-matched dSYMs, captures one launch or runtime trace, and preserves processed output JSON before analysis. It installs the ettrace Homebrew runner, builds simulator frameworks from upstream ETTrace when needed, and gates analysis on complete symbolication for app-owned binaries. Launch traces use ettrace --simulator --launch while runtime flows perform one focused interaction then stop the runner. Processed output_<thread>.json files are copied immediately because later runs overwrite them, and analysis uses bundled analyze_flamegraph_json.py on fresh artifacts only. Reports include flow description, simulator details, top first-party stacks, symbol completeness, and caveats about simulator-only variance. Cleanup removes temporary ETTrace wiring unless the user requests otherwise. Use when profiling iOS Simulator launch or runtime latency with ETTrace flamegraphs.45installs7Frontend Testing DebuggingThe frontend-testing-debugging skill tests, debugs, and improves rendered frontend apps through the Build Web Apps or web dev plugin workflow. It infers target surfaces from repo scripts, open URLs, or running dev servers, then runs a validation loop: identify flow, choose browser path, make smallest edit, validate rendered behavior, and reply with a QA report. Browser availability is classified as available, absent, or invocation failed, preferring the Browser plugin skill over regular Playwright when listed. Required browser checks cover page identity, non-blank DOM, no framework error overlay, console health, screenshot evidence, and interaction proof. Regular Playwright fallback records the exact Browser failure reason. Target flows are stated as entry route, user action, and expected rendered result. Visual work adds desktop and mobile viewports with a mismatch ledger for reference-driven fixes.28installs8Liquid GlassThe liquid-glass skill modernizes macOS SwiftUI apps to use system Liquid Glass materials with minimal custom chrome. It audits NavigationSplitView, TabView, sheets, toolbars, and floating controls, removing custom backgrounds and scrims that obscure system glass. Workflow updates standard structure first, then adds custom glassEffect only where standard controls do not suffice. Toolbar guidance covers ToolbarSpacer grouping, sharedBackgroundVisibility for standalone items, badge indicators, and scrollEdgeEffectStyle instead of custom bar backgrounds. Search uses searchable on NavigationSplitView with searchToolbarBehavior for compact affordances. Sheets should drop presentationBackground imitations and use navigation zoom transitions from toolbar sources. For SwiftPM GUI apps, pair with build-run-debug to launch foreground .app bundles. Validates glass grouping, transitions, icon treatment, and keyboard usability.27installs9Build Run DebugThe build-run-debug skill sets up project-local script/build_and_run.sh as the single kill, build, and run entrypoint for macOS apps, wiring .codex/environments/environment.toml for a Codex Run button. It discovers git repo state, .xcworkspace, .xcodeproj, or Package.swift shapes, resolves runnable schemes or SwiftPM executable products, and distinguishes command-line tools from AppKit or SwiftUI GUI apps. GUI apps get a generated dist/AppName.app bundle with Info.plist and open -n launch instead of raw swift run. Optional flags add lldb debug, log stream, telemetry subsystem filtering, and pgrep process verification. Failures are classified as compiler, linker, signing, build settings, SDK, script, or runtime launch issues with minimal useful error snippets. SwiftPM GUI apps must not use direct executable launch. References run-button-bootstrap.md as canonical for script and environment format.25installs10Packaging NotarizationThe packaging-notarization skill addresses shipping macOS apps beyond local debug runs: archives, exported bundles, hardened runtime, and notarization readiness. Workflow confirms distribution goals such as local archive validation, signed distributable, or notarization troubleshooting, then inspects app bundle structure including nested frameworks, helper tools, and entitlements. It reviews hardened runtime, signing identity, nested code signatures, and required entitlements, separating packaging issues from trust-policy symptoms. Guardrails avoid presenting notarization as required for ordinary debug runs and flag when conclusions are inferred without the actual exported artifact. Output states what was inspected, distribution readiness, top missing prerequisite, and next validation command. Pair with signing-entitlements for Gatekeeper failures.25installs11Signing EntitlementsThe signing-entitlements skill diagnoses macOS failures that smell like codesigning rather than compilation: launch refusal, missing entitlements, invalid signatures, sandbox mismatches, hardened runtime confusion, or trust-policy rejection. It locates the .app or executable, reads signing with codesign -dvvv --entitlements and spctl -a -vv for Gatekeeper, and inspects plists with plutil -p. Failures are classified as unsigned, wrong identity, entitlement mismatch, hardened runtime, App Sandbox, nested signing, or distribution prerequisite issues. The minimum fix path quotes exact problems and shortest validation commands, distinguishing local development from distribution problems. Useful commands include security find-identity for signing identities. Guardrails forbid inventing entitlements and avoid conflating notarization with debug signing.25installs12Swiftpm MacosThe swiftpm-macos skill handles package-first macOS repositories where Package.swift is the main entrypoint or faster than opening Xcode. It reads Package.swift to identify executable, library, and test products, runs swift build by default and release mode only on request, and uses swift run for executables explaining defaults when multiple exist. Testing uses swift test with optional filters for specific targets or cases. Failures are summarized as module resolution, package graph, linker, runtime, or test regression blockers. Guardrails prefer SwiftPM when simpler than Xcode, avoid assuming app bundles in pure packages, and note library-only products are not directly runnable. Output reports discovered products, commands run, success state, and top blocker. Pair with build-run-debug when GUI apps need .app bundle staging.23installs13Swiftui PatternsThe swiftui-patterns skill builds macOS SwiftUI scenes and components using desktop-native patterns for windows, commands, toolbars, settings, split views, and inspectors. It prefers system affordances such as commands, toolbars, sidebars, contextual menus, and searchable over custom chrome. Workflow chooses ownership location first, then wrappers, avoiding reflex view models for simple desktop state. Guidance covers NavigationSplitView column visibility, inspector placement, settings scenes, menu command organization, and toolbar item grouping with shared backgrounds. It pairs with view-refactor and window-management skills for large view splits and chrome tuning. Validation checks keyboard navigation, selection persistence, and scene restoration behavior. Use when shaping new macOS UI surfaces or aligning existing views with current desktop HIG patterns. Identify the feature or scene and the primary interaction model: document, editor, sidebar-detail, utility window, settings, or menu bar extra.23installs14View RefactorThe view-refactor skill refactors macOS SwiftUI views and scenes into stable structure when files grow unwieldy or AppKit bridges proliferate. It splits giant root views that silently own every desktop surface, preferring one stable split or window layout with local conditionals inside. Extraction rules keep scene-level state at the scene root, push visual subtrees into focused child views, and isolate AppKit representables behind narrow adapters. It avoids duplicating state between SwiftUI and AppKit layers and documents when AppKit escape hatches are justified versus replaceable with SwiftUI modifiers. Pair with swiftui-patterns for desktop affordances and window-management for chrome behavior. Output describes before structure, extracted components, state ownership map, and validation steps for navigation and restoration. only at the narrow edges where desktop behavior truly requires it. Break the app into meaningful scene roots: main window, settings, utility windows, inspectors, or menu bar extras.23installs15Window ManagementThe window-management skill customizes macOS SwiftUI windows and scene behavior when tuning chrome, drag regions, placement, restoration, launch behavior, or borderless windows. It prefers scene and window modifiers over ad hoc AppKit bridges when SwiftUI offers equivalents, and system materials over hardcoded translucent fills for stylized windows. Workflow inspects current scene declarations, identifies chrome goals such as unified toolbar, hidden title bar, or draggable regions, then applies windowStyle, windowToolbarStyle, defaultPosition, and restoration behaviors. Borderless and panel-like windows document focus and key-window implications. Pair with build-run-debug for launch validation and appkit-interop only when SwiftUI lacks the needed NSWindow hook. Output states modifiers applied, AppKit bridges if any, and manual QA steps for resize, fullscreen, and multi-window restoration. restoration behavior, and initial or zoomed placement. resizability, restoration, and placement rules. deployment targets, expect to use more AppKit bridging or availability guards.23installs16Appkit InteropThe appkit-interop skill bridges macOS SwiftUI into AppKit narrowly when implementing representables, reaching NSWindow or panels, handling menus, or using the responder chain. It avoids duplicating source of truth between SwiftUI and AppKit state, keeps Coordinator objects focused, and limits bridge surface to APIs SwiftUI cannot express. Workflow identifies the minimal NSView or NSViewController representable, documents lifecycle callbacks, and validates focus, key window, and menu validation behavior. Pair with swiftui-patterns and window-management before reaching for AppKit. Output describes bridge scope, state ownership, and manual QA steps for keyboard shortcuts and menu items. Guardrails forbid letting Coordinator become an unstructured dumping ground or using AppKit when SwiftUI modifiers suffice. Use this skill when SwiftUI is close but not quite enough for native macOS behavior. the source of truth, while AppKit handles the imperative edge.22installs17TelemetryThe telemetry skill adds and verifies lightweight macOS runtime telemetry using Logger from OSLog without turning the codebase into a logging landfill. It prefers structured Logger categories for windows, sidebars, menus, and user actions, using privacy annotations appropriate for release builds. Workflow identifies high-signal events, adds concise log points at state transitions, and verifies output via Console or log stream filters during manual QA. Pair with build-run-debug for launch scripts that capture subsystem logs. Output lists added categories, sample log lines, and verification commands. Guardrails avoid NSLog spam, logging secrets, or per-frame logging in hot paths. logging APIs and verify the events after a build/run loop. Give each feature a clear subsystem/category pair so runtime filtering stays easy. Log meaningful user and app lifecycle events: window opening, sidebar selection changes, menu commands, menu bar extra actions, sync/load milestones, and unexpected fallback paths.22installs18Shadcnshadcn is a Claude Code skill for ai & agent building. It helps solo builders move faster with AI-assisted coding.21installs19Test TriageThe test-triage skill narrows macOS test failures without treating every red test as a product bug. It detects whether xcodebuild test or swift test is appropriate, runs the smallest meaningful scope using targets or filters when provided, and classifies outcomes as build failure, assertion failure, crash, async flake, environment setup, or entitlement issues. Focused reruns avoid full-suite loops without new signal. Output reports the command, smallest failing scope, top failure category, likely cause, and next proof step. Guardrails separate compilation from execution failures, flag iOS-only assumptions on macOS, and mark probable flakes without overstating confidence. Detects Xcode versus SwiftPM harness automatically. Runs smallest failing target or filter first. Classifies build, assertion, crash, flake, and setup failures. Recommends focused reruns over full-suite loops. Summarizes command, scope, category, and next step. Clear failure category with command, scope, and next rerun step.21installs20React Best PracticesThe react-best-practices skill applies Vercel Engineering performance guidance across React and Next.js work. It organizes 64 rules into eight impact-prioritized categories from eliminating async waterfalls and bundle size through server-side caching, client fetching, re-render control, rendering performance, JavaScript micro-optimizations, and advanced patterns. Each rule uses a prefixed identifier such as async-parallel or bundle-dynamic-imports with dedicated markdown files showing incorrect and correct examples. Use when implementing components, data fetching, bundle trimming, or performance reviews. The skill points to individual rule files and the compiled AGENTS.md for full detail. Priority tables steer agents toward CRITICAL waterfall and bundle fixes before lower-impact rendering tweaks. 64 rules across eight impact-prioritized categories. Covers waterfalls, bundles, server cache, and re-renders. Each rule file shows incorrect and correct examples. Triggers on React, Next.js, and performance tasks. References compiled AGENTS.md for the full guide.19installs21Game Ui FrontendThe game-ui-frontend skill guides UI surfaces for browser games where canvas or WebGL renders the world and DOM handles text-heavy chrome. Workflow establishes visual direction first across genre fantasy, materials, typography, palette, and motion tone, then implements CSS variable themes with hierarchy placing combat or survival data first and settings behind menus. It protects the playfield in 3D by limiting persistent HUD clusters, keeping the center clear, and deferring dense panels. Standards require readable thematic interfaces rather than generic dashboards, with responsive layouts that scale overlays without obscuring gameplay. Use when users request HUDs, menus, overlays, responsive game UI, or visual direction tied to play experience.12installs22Stripe Best PracticesThe stripe-best-practices skill routes integration questions using API version 2026-02-25.clover and a routing table mapping one-time payments to Checkout Sessions, embedded Payment Element flows to Checkout plus Payment Element, saved payment methods to Setup Intents, Connect marketplaces to Accounts v2, subscriptions to Billing plus Checkout, and embedded banking to v2 Financial Accounts. It instructs reading references/payments.md, connect.md, billing.md, and treasury.md before answering or coding. Supplemental docs include Integration Options, API Tour, and Go Live Checklist. Use when building, modifying, or reviewing any Stripe surface including marketplaces, subscriptions, or Treasury accounts.11installs23Game Studiogame-studio routes early browser-game work across 2D Phaser and 3D Three.js or React Three Fiber paths. Classifies requests, locks core loop and stack, then hands off to specialist skills for implementation, assets, UI, and playtesting.8installs24Phaser 2d Gamephaser-2d-game skill documents Implement 2D browser games with Phaser. Use when the user wants a Phaser, TypeScript, and Vite stack for scenes, gameplay systems, cameras, sprite animation, and DOM-overlay HUD patterns.. name: phaser-2d-game description: Implement 2D browser games with Phaser. Use when the user wants a Phaser, TypeScript, and Vite stack for scenes, gameplay systems, cameras, sprite animation, and DOM-overlay HUD patterns.8installs25Supabase Postgres Best Practicessupabase-postgres-best-practices is a Claude Code skill for databases. It helps solo builders move faster with AI-assisted coding.8installs26Web Game Foundationsweb-game-foundations sets browser-game architecture before coding. Separates simulation from rendering, defines input action maps, asset manifest policy, save/debug boundaries, and engine selection between Phaser, Three.js, and React Three Fiber.8installs27Game Playtestgame-playtest skill documents Run browser-game playtests and frontend QA. Use when the user asks for smoke tests, screenshot-based verification, browser automation, HUD or overlay review, or structured issue-finding in a browser game.. name: game-playtest description: Run browser-game playtests and frontend QA. Use when the user asks for smoke tests, screenshot-based verification, browser automation, HUD or overlay review, or structured issue-finding in a browser game.6installs28Openai Platform Api Keyopenai-platform-api-key skill documents Use when Codex is asked to build, run, test, debug, or configure an OpenAI-backed or provider-unspecified AI app, UI, script, CLI, generator, or tool, especially requests phrased only as "using AI" or generators driven by forms/user input; also use for OPENAI_API_KEY or sk-proj setup. Treat this as . name: openai-platform-api-key description: 'Use when Codex is asked to build, run, test, debug, or configure an OpenAI-backed or provider-unspecified AI app, UI, script, CLI, generator, or tool, especially requests phrased only as "using AI" or generators driven by forms/user input; also use for OPENAI_API_KEY or sk-proj setup. Treat this as the credential gate: inspect safely, ask reuse-vs-new be6installs29Openai Api Troubleshootingopenai-api-troubleshooting skill documents Use when an OpenAI API request fails and Codex needs to classify the likely cause, explain the next step, and route to the right follow-up. Covers common runtime failures such as blocked outbound network access, invalid credentials, exhausted API quota or credits, rate limits, and model, project, or. name: openai-api-troubleshooting description: Use when an OpenAI API request fails and Codex needs to classify the likely cause, explain the next step, and route to the right follow-up. Covers common runtime failures such as blocked outbound network access, invalid credentials, exhausted API quota or credits, rate limits, and model, project, or organization access issues; delegate key provisioning5installs30Research Router Skillresearch-router-skill skill documents Route broad or ambiguous life-sciences research requests to the right skills, normalize core entities, optionally parallelize independent evidence gathering with subagents when available, and synthesize a concise evidence-backed answer. Use when a user asks a general life-sciences question that coul. name: research-router-skill description: Route broad or ambiguous life-sciences research requests to the right skills, normalize core entities, optionally parallelize independent evidence gathering with subagents when available, and synthesize a concise evidence-backed answer. Use when a user asks a general life-sciences question that could span multiple sources or analysis types.5installs31Sprite Pipelinesprite-pipeline skill documents Generate and normalize 2D sprite animations. Use when the user asks for full-strip generation from approved source frames, consistent anchor and scale normalization, or preview assets for browser-game animation.. name: sprite-pipeline description: Generate and normalize 2D sprite animations. Use when the user asks for full-strip generation from approved source frames, consistent anchor and scale normalization, or preview assets for browser-game animation.5installs32Ipd Skillipd-skill skill documents Submit compact IPD REST requests for HLA allele and cell-level metadata using the public IPD query API. Use when a user wants concise IPD summaries; save raw JSON or text only on request.. name: ipd-skill description: Submit compact IPD REST requests for HLA allele and cell-level metadata using the public IPD query API. Use when a user wants concise IPD summaries; save raw JSON or text only on request.4installs33Ncbi Pmc Skillncbi-pmc-skill skill documents Submit compact NCBI PMC Open Access requests for article/file availability metadata. Use when a user wants concise PMC Open Access summaries; save raw XML only on request.. name: ncbi-pmc-skill description: Submit compact NCBI PMC Open Access requests for article/file availability metadata. Use when a user wants concise PMC Open Access summaries; save raw XML only on request.4installs34Verificationverification skill documents Full-story verification - infers what the user is building, then verifies the complete flow end-to-end: browser → API → data → response. Triggers on dev server start and 'why isn't this working' signals.. name: verification description: "Full-story verification - infers what the user is building, then verifies the complete flow end-to-end: browser → API → data → response. Triggers on dev server start and 'why isn't this working' signals."3installs35Authauth is a Claude Code skill for security. It helps developers move faster with AI-assisted coding.2installs36Search Company Knowledgesearch-company-knowledge skill documents Search across company knowledge bases (Confluence, Jira, internal docs) to find and explain internal concepts, processes, and technical details. When an agent needs to: (1) Find or search for information about systems, terminology, processes, deployment, authentication, infrastructure, architecture,. name: search-company-knowledge description: "Search across company knowledge bases (Confluence, Jira, internal docs) to find and explain internal concepts, processes, and technical details. When an agent needs to: (1) Find or search for information about systems, terminology, processes, deployment, authentication, infrastructure, architecture, or technical concepts, (2) Search internal documentati2installs37Figma Generate Designfigma-generate-design skill documents Use this skill alongside figma-use when the task involves translating an application page, view, or multi-section layout into Figma. Triggers: 'write to Figma', 'create in Figma from code', 'push page to Figma', 'take this app/page and build it in Figma', 'create a screen', 'build a landing page in . name: figma-generate-design description: "Use this skill alongside figma-use when the task involves translating an application page, view, or multi-section layout into Figma. Triggers: 'write to Figma', 'create in Figma from code', 'push page to Figma', 'take this app/page and build it in Figma', 'create a screen', 'build a landing page in Figma', 'update the Figma screen to match code'. This is t1installs38Notion Knowledge Capturenotion-knowledge-capture skill documents Capture conversations and decisions into structured Notion pages; use when turning chats/notes into wiki entries, how-tos, decisions, or FAQs with proper linking.. name: notion-knowledge-capture description: Capture conversations and decisions into structured Notion pages; use when turning chats/notes into wiki entries, how-tos, decisions, or FAQs with proper linking.1installs