Claude Skills
The open format is called Agent Skills and works in Claude Code, Codex, Cursor and other agents — most people know it as Claude Skills.
Every Agent Skill we could find on GitHub, deduplicated by content. 79 600 files from 1 763 authors, of which 61 947 are unique — the rest is the same skill repackaged into someone else's repository. For each one: what it weighs in tokens, whether it ships runnable scripts, and which MCP servers it needs.
Analyzes git repositories and commit history
Per-case mocked MCP override smoke skill used by skill-up e2e tests.
MCP stdio marker smoke skill used by skill-up e2e tests.
Verify Sandbox MCP can be installed and called from skill-up.
SDD-RIPER workflow executor with phase gating
A skill with potentially flaky behavior for testing retry
Triggered when the user needs to convert data formats, such as CSV to JSON, JSON to YAML, or XML to JSON. Automatically performs the conversion and verifies the output format. Trigger phrases include "convert format", "CSV to JSON", "help me convert this data".
Triggered when the user submits code or requests a code review. Automatically analyzes code quality, identifies potential bugs, security vulnerabilities, and performance issues, and provides improvement suggestions. Trigger phrases include "take a look at this code", "review this", "is there a problem with this function".
Triggered when the user expresses boredom, low mood, frustration, irritation, or a need to relax. Responds with warm words of comfort and tells a joke to lighten the atmosphere. Trigger phrases include "I'm so bored", "feeling down", "so frustrated", "gloomy", "not happy", "I'm tired".
Create, run, diagnose, and iteratively improve Agent Skill evaluations (evals) with the skill-up CLI / 使用 skill-up CLI 创建、运行、诊断并持续改进 Agent Skill 评测. Use when the user asks to evaluate, test, regress, verify, fix, improve, iterate, or evolve a Skill; add or strengthen eval cases; write eval.yaml/case.yaml; run skill-up run/validate/list-cases/report/import/init; or migrate from Anthropic evals.json. Handles Skill discovery, eval scaffolding, judge authoring, validation, runs, reports, and evidence-based repair loops.
Validate, run, and execute workflows; creating new workflows when explicitly asked. Use when orchestrating AI agents via YAML workflow files, executing an existing workflow, debugging execution, configuring routing between agents, setting up human-in-the-loop gates, or understanding workflow YAML schema. Only create new workflows when explicitly asked.
>- GIF/WebP animations through the local one-file image-use CLI (formerly chatgpt-imagegen), using the user's ChatGPT subscription by default, the Codex backend as fallback, or an optional Gemini subscription — no API key 画图, 画一张, 生成图片, 生图, 配图. Use for photos, illustrations, icons, hero banners, mockups, sprites, concept art, animation loops, and figures for documents, proposals, blog posts, or READMEs; save outputs in the workspace. Auto mode prefers the logged-in ChatGPT browser through chrome-use to avoid Codex usage and falls back to the Codex backend only when the web path is unavailable. Users with a Gemini subscription can name --backend gemini or agy instead. Proactively propose useful figures while authoring long-form content. Do not use for editing existing images, SVG/vector work, code-native graphics, established icon systems, explicit high-quality or transparent API output, or end-user image-generation services.
Test environment system (authored DomeGradient and IBLGradient, component schemas) against the poke example using the iwsdk CLI.
Create, place, preview, debug, and connect manifest-backed IWSDK UIKitML assets. Use for spatial panels, browser HUDs, UIKitML layout, fonts, editor previews, or runtime element behavior.
Test grab system (distance grab, one-hand grab, two-hand grab) against the grab example using the iwsdk CLI.
Test ECS core functionality (system registration, components, Transform sync, pause/step/resume, system toggle, entity discovery, snapshots) against the poke example using the iwsdk CLI.
IWSDK experience pipeline and planning guide. Use when building a new IWSDK app/game end-to-end, planning new IWSDK features, designing systems/components, reviewing IWSDK code architecture, or when the user asks about IWSDK patterns, ECS design, signals, or reactive programming. Runs a phased ideation → design → grounding → architecture → build → verify → ship pipeline, orchestrating sub-agents per phase where the harness supports them.
Parallel test orchestrator. Runs all 9 test suites concurrently via Task sub-agents and the iwsdk CLI. Handles build, example setup, dev servers, agent launch, polling, retries, and result aggregation.
Find and click a target object in XR. Use when testing UI interactions, clicking buttons, or verifying interactable elements work correctly.
Test audio system (AudioSource loading, playback state, stop, spatial audio) against the audio example using the iwsdk CLI.
Test locomotion system (slide, snap turn, teleport, jump) against the locomotion example using the iwsdk CLI.
Test XR session lifecycle and mode transitions. Use when verifying XR enter/exit behavior, testing mode-dependent features, or debugging session state issues.
Test level system (LevelRoot, LevelTag, authored environment, scene hierarchy) against the poke example using the iwsdk CLI.
Debug continuous behavior in WebXR scenes — physics, animations, collisions, game loops, or any real-time interaction that happens too fast for an agent to observe. Uses ECS pause/step/snapshot/diff to freeze time and inspect state frame by frame.
Test Havok physics system (gravity, rigid bodies, static vs dynamic) against the physics example using the iwsdk CLI.
Test UIKitML asset and ScreenSpace UI behavior against the poke example using the iwsdk CLI.
Migrate an existing IWSDK 0.4.x application to IWSDK 0.5.0. Use when upgrading across the 0.5 boundary, replacing Meta Spatial Editor or GLXF, removing retired Vite plugins, moving UIKitML to runtime-loaded assets, adopting native scene JSON, or resolving 0.5 migration errors.
Test XR interactions (ray, poke/touch, dual-mode, audio, UI panel) against the poke example using the iwsdk CLI.
Guide for implementing depth sensing and occlusion in IWSDK projects. Use when adding depth-based occlusion to hide virtual objects behind real-world surfaces, configuring DepthSensingSystem, choosing occlusion modes, or troubleshooting objects that disappear or fail to occlude.
Grab an object in the WebXR scene using emulated controllers. Use when the user wants to pick up, move, or test grabbing an object. Supports OneHandGrabbable and TwoHandsGrabbable components which use proximity-based grip (squeeze button), not trigger.
Ray-based interactions in the WebXR scene — click objects, press UI buttons, or distance-grab with DistanceGrabbable. Use when the user wants to point at and interact with something at a distance, click a UI button, or test ray-based selection.
Guide for implementing physics in IWSDK projects. Use when adding physics simulation, configuring rigid bodies, collision shapes, applying forces, creating grabbable physics objects, or troubleshooting physics behavior.
Compose editable static IWSDK scenes from text, images, or hybrid references using a shared application asset manifest, v1 scene JSON, modular scene files, and the managed editor's validation and rendering tools. Use for 3D environments, props, architecture, staged scenes, procedural Three.js assets, custom PBR or shader materials, glTF assets, prefabs, patterns, lighting, camera matching, or visual review.
Create, place, preview, debug, and connect manifest-backed IWSDK UIKitML assets. Use for spatial panels, browser HUDs, UIKitML layout, fonts, editor previews, or runtime element behavior.
Searches, saves, and maintains a local document index through a local RAG MCP server. Use when user says "search my docs", "save this page", "read around that chunk", "sync my index", or invokes `npx mcp-local-rag`.
Use before handing back any long-form draft: edit in separate passes for sense, voice, reader payoff, proof, and specificity rather than re-reading and tweaking.
Use when building, editing, or sending anything in Resend, including broadcasts, templates, segments, and contacts, or when checking what a past send did.
Use when turning existing copy, a blog post, an announcement, or a brief into an email that works in an inbox, including deciding what to cut and what to link out to.
Use before writing or updating the team's shared brand context document: what belongs in it, how it's structured, how to merge new information without losing what's there, and what to keep out.
Use when drafting or editing a blog post or long-form article in the house voice.
Use when writing or editing the copy of an email, newsletter, or broadcast in the house voice, including subject lines and preview text.
Use when the question is whether email will reach the inbox, covering sending domain setup, authentication, list hygiene, complaint and bounce rates, or a drop in delivery.
Use when deciding what content to write: matching topics to buyer stages, choosing between searchable and shareable pieces, and grouping topics into pillars and clusters.
Use when planning many similar pages from a template and a data set: choosing the pattern, deciding whether the data can carry it, and avoiding the thin-content failure that gets page sets deindexed.
Use when you need to find out what customers actually think rather than assume it: what to ask in an interview, how to mine language from reviews and support conversations, and how to tell a real finding from a polite one.
Use when drafting or editing a post for Bluesky.
Use when turning settled positioning into the lines other people will reuse: the message hierarchy, value propositions per audience, objection handling, and the proof behind each claim.
Use when adding, fixing, or reviewing structured data: JSON-LD for articles, products, FAQs, breadcrumbs, organizations, and local businesses, and what to do when markup doesn't earn a rich result.
Use when reviewing a page or site for organic search, or diagnosing why something isn't ranking: crawl and index signals, on-page elements, content quality, and how to order the fixes.
Use when working out what a product is, who it's for, and what it competes with: the market alternatives, the ideal customer, the differentiators worth claiming, and the positioning statement that follows.
Use when planning or restructuring what pages a site has and how they connect: hierarchy, navigation, URL patterns, breadcrumbs, and internal linking. Not for XML sitemaps, which are in seo-audit.
Use when drafting or editing a LinkedIn post in the house voice.
Use when drafting or editing a post for Threads (Meta): casual, conversational, reply-driven voice.
Use when drafting or editing a post or thread for X (Twitter): voice, hooks, length, threads, hashtags, and links.
Use when drafting or editing a post (toot) for Mastodon / the fediverse.
Draft release notes from merged PRs, propose a semver bump, and emit a copy-pasteable `gh release create` command. Use when preparing a tagged GitHub release.
> Auto-generate release notes from a list of PRs for Azure Functions Python Worker components (worker, runtime v1, or runtime v2). Analyzes PR file changes to determine component relevance, groups by category/prefix, and outputs organized release notes. Use when the user asks to "generate release notes", "create release notes from PRs", "format these PRs for release", or provides a list of PRs to organize.
> Use when the user wants to design experiments, plan ablation studies, structure baselines, or create incremental evaluation strategies. Triggers on phrases like "design ablation", "plan experiment", "what experiments should I run", "baseline comparison", or "experiment matrix".
Claude Skills — questions
Answers built from the skills we actually parsed.