Skills
100,696 packaged capabilities and playbooks for agents. Install any of them with one command.
All
React Router v7 (Framework Mode, formerly Remix) reference. loader, action, middleware, hooks (useLoaderData / useNavigate / useFetcher / useNavigation, etc.), components (Link, NavLink, Form, Outlet), routes.ts, SSR / SPA / Static, ErrorBoundary, session, redirect, defer, type-safe routing.
Build UIs with React Spectrum S2 (Spectrum 2), Adobe's component library for React. Use when developers are using `@react-spectrum/s2` or need Adobe design system components. Includes React Aria Components docs as a reference for building custom components on top of unstyled primitives.
Post to Reddit using AppleScript Chrome control. Use when user wants to post to Reddit, share on subreddits, or promote open source projects on Reddit. Triggers on "post to reddit", "share on reddit", "reddit post", "submit to subreddit", or any Reddit posting request.
Rebuilds the current implementation from scratch as the simplest design that satisfies what was learned, after the user approves the design. Use when the user explicitly asks to redo the work, start over, scrap the approach, or rewrite it more simply. Never infer from a failing or complex implementation. Not for fitting a new requirement into an existing design; use redesigning-from-first-principles.
Removes every source comment outside a short keep list, flagged by a fresh-context reviewer, and, with approval, fixes root causes and encodes enforceable constraints in types, tests, or lint. Use when the user explicitly asks to remove or clean up code comments. Never infer from a diff containing comments.
This skill should be used when the user asks to implement, optimize, debug, or review TUI rendering, layout, repainting, flicker, frame buffers, diff rendering, terminal resize behavior, alternate screen lifecycle, scrollback, viewports, tables, dashboards, progress displays, or performance over SSH/tmux. PROACTIVELY activate for: rendering performance, layout bugs, flicker, alternate screen cleanup, resize handling, terminal dashboards, frame pacing, virtual DOM/cell buffers, double buffering, dirty regions, scroll regions, long-running tasks, and high-frequency updates. Provides: render-loop rules, layout patterns, performance budgets, lifecycle cleanup checks, and failure triage.
Ingest .mp4 video into Rerun chunk streams with rerun.experimental.Mp4Reader. Read when converting video into a VideoStream, choosing stream vs asset mode, transcoding (B-frames, output codec, GOP size) through FFmpeg, or aligning video PTS onto a recording's wall-clock timeline. Builds on rerun-chunk-processing (stream mechanics) and rerun-data-model (where the video belongs in the recording).
Run fast, source-grounded UX research on the highest-signal problems users are experiencing with a user-specified digital product. Use when the user asks to research user pain, UX friction, onboarding issues, docs/help problems, developer experience friction, support pain, product workflow issues, or current user complaints for a named product.
Reverse engineer React Native mobile apps, including Hermes bytecode bundles, using hbctool, hermes-dec, and Frida. Use when an APK contains index.android.bundle or libhermes.so, when an IPA contains main.jsbundle, when jadx shows only ReactActivity classes, or when a bundle file starts with the Hermes magic bytes instead of readable JavaScript.
Power-user audit of Origin's pending surfaces. Most users want `/brief` for revisions. That handles the daily flow. Use `/review` only for explicit deep-walk audits after bulk imports, or when you want to walk the full queue rather than the top 3 shown in /brief. Invoked as `/review captures` or `/review revisions`.
Reviews a diff, branch, pull request, or named files with a panel of parallel reviewer models, then merges and triages their findings and optionally applies approved fixes. Use only when the user asks for a panel, multi-model, or multi-agent review, since it spawns several agent sessions; for an ordinary review use code-review.
Review Supabase docs changes locally in your `supabase/supabase` checkout — either an open PR (triage, classify, verify) or your own branch before opening a PR (local self-review). Covers markdown pipeline, MDX content, tutorials, examples, Studio links, and docs tooling. Use when asked to review docs PRs, self-review a draft branch, check who has approved, verify build output, or evaluate supabase/supabase documentation changes.
Reviews a code diff (PR, branch, commit range, working tree, or a diff already in the conversation) in a fresh-context read-only subagent, because a session holding the conversation is a biased reviewer, posts its Conventional Comments findings inline and its APPROVE, REQUEST CHANGES, or COMMENT verdict to the pull request when one exists, and returns a short summary. Use when the user says "code review", or when asked to review code, a diff, or a PR, including a pasted diff, a .diff or .patch file, or a diff shown earlier; never review it inline. Not for design docs; use reviewing-design-docs. Not for a security-focused review; use reviewing-security.
Perform RFM (Recency, Frequency, Monetary) customer segmentation analysis on e-commerce data. Use when you need to analyze customer value, identify VIP customers, or create marketing segments. Automatically cleans data, calculates RFM metrics, applies K-means clustering, and generates visualization reports with Chinese language support.
Source, verify and wire the sound for a Roblox game from the Creator Store by asset id - footsteps, gunshots, animal calls, UI clicks, ambience, music - with no upload, no Open Cloud audio quota and no API key. Searches the same unauthenticated toolbox endpoints Studio's own audio tab uses, then PROVES each candidate in Studio (IsLoaded plus a measured Sound.TimeLength) before anything is chosen, because the toolbox's own duration field is wrong often enough to be dangerous. Emits a SoundBank data module, an Sfx runtime helper that fails inert, and an install script so the bank can be heard in an open place immediately. Use whenever the user wants sound effects, audio, SFX, music or ambience for a Roblox game, wants sounds found and wired rather than dragged in by hand, is picking audio for a build, or asks why a Sound is silent, has no length, or will not load in their place.
Use when anything behaves unexpectedly — a failing test, an error or exception, a crash, a reported bug, wrong output, a performance regression, a flaky CI job — and before proposing or applying any fix. Produces an evidence-backed investigation whose authoritative causal confirmation is a human disposition. Not for a failure that is only on a deployed environment with no pack yet (debug-remote).
Use Roundfix to plan releases with the read-only Release Plan Command, clean CodeRabbit pull request feedback, diagnose runtime readiness with the Doctor Command, execute a Spec's Task Graph with the Implement Command, monitor Runs through the Supervisor Run Event Stream, reclaim Run storage with the GC Command, archive completed Specs, run its `--follow-up`, `--status` and `--unattended` Skill Modes, and, inside daemon-assigned Batch runs, follow the bounded Review Issue or Task resolution contract.
Ruby on Rails architecture and coding patterns from Basecamp. Use when writing, reviewing, or refactoring Rails code to follow proven conventions for models, controllers, jobs, and concerns. Triggers on tasks involving Rails models, concerns, controllers, background jobs, or Turbo/Hotwire.
Investigate why a Duvo Run failed or produced the wrong outcome. Use when the user shares a failed Run, asks "why did this Run fail", or wants to fix a recurring failure on an Agent. Reads the Run's transcript and the Build that was active for it via the Duvo public API, names the root cause from a fixed failure-mode taxonomy, and proposes one concrete fix — handing off to aop-writer for any AOP rewrite.
Use when designing, implementing, testing, or debugging Rust service workflows that must be safe under retries, duplicate requests, crashes, concurrent submissions, background workers, queues, email/payment/notification side effects, idempotency keys, transaction isolation, or save-and-replay API behavior.
Make a promo / motion video for a software product with HyperFrames, using the saas-motion-kit 7-stage process (brief → components → theme → storyboard → build → sound → deliver), the clean-or-imaginary component rule, the creative muscle (tone matrix, variety rules, component forge, transition atlas, variety audit) and the 101-theme gallery in docs/. Use when someone asks for a launch video, product promo, booth loop, feature reveal or social motion piece for an app, SaaS or developer tool.
SAFT (Simple Agreement for Future Tokens) review and advisory skill for crypto founders, token issuers, and investors. Use when user (1) uploads a SAFT/token warrant for review, (2) asks questions about how SAFTs work, (3) requests to draft a SAFT, or (4) asks about token fundraising structures. Triggers on keywords like SAFT, token warrant, token side letter, future tokens, token sale, ICO, TGE, token generation event, vesting, lockup, crypto fundraising.
Uses webinars to sell — live and automated/evergreen webinar strategy, presentation structure, registration funnels, follow-up sequences, and conversion optimization. Use when you're not sure how to structure a selling webinar, your webinar attendance rate is low, attendees aren't converting into buyers, your evergreen webinar performs worse than live, or you can't decide which webinar platform to use. Do NOT use for general funnel strategy (use /sales-funnel), email follow-up sequences (use /sales-email-marketing), or general sales call prep (use /sales-call-review). For Groove-specific help, use /sales-groove.
Create polished audio content and save to Spotify. Produces episodes with TTS narration, a rich timeline (chapters plus in-player images, external links, and Spotify entity cards), and a cover image. Also use for raw media saves, show/episode management, and timeline navigation.
Transform a user-supplied photograph into an original vertical 3:4 art-directed poster using watercolor, pop-art screenprint, expressive painting, ink-wash, relief print, editorial surrealism, or a controlled hybrid. Use when the user asks to reinterpret a real scene, portrait, object, vehicle, landscape, architecture, or travel photo as creative visual art, an exhibition poster, editorial poster, key visual, or unified poster series with source-specific typography, editorial micro-copy, image-title interlocking, and selected identity preservation.
The academic paper writing and polishing skill in the SciPilot Skills family, positioned to put a "top-journal editor + harsh reviewer" into your writing workflow. It covers the full chain of writing tasks: Chinese-English mutual translation and polishing, condensing/expanding, English/Chinese expression polishing, logic checking, removing AI flavor (humanize), chapter-by-chapter drafting (Title/Abstract/Introduction-CARS/Methods/Results/Discussion/Conclusion/Limitations), figure/table caption generation, experimental results analysis, self-check from a reviewer's perspective, cover letter, response-to-reviewers (rebuttal). Unlike similar "prompt checklists", this skill comes with a [writing quality evidence chain]: before every delivery it mandatorily runs scripts/writing_lint.py for machine self-checking (AI fingerprint words, mechanical connectives, dangling -ing, dash overuse, negative parallelism, vague attribution, model-name possessives, LaTeX special character escaping, Markdown residue in Word, Chinese full-width punctuation, passive ratio, sentence length rhythm), then adds an AI read-through reviewer-perspective self-review loop (logic/exaggeration/AI flavor/consistency → revise, up to 3 rounds). Academic integrity red lines: polishing changes only expression, not values/direction/conclusions; never fabricate data or citations; the strength of a claim must match the strength of the evidence. Stage 0 must proactively ask item by item: task type, text carrier (.tex/.docx/.md/plain text), target journal or conference, language direction, discipline, revision conservativeness; silently assuming is forbidden. Trigger this skill when the user mentions any of the following: 润色, polish, proofread, 改写, 润色论文, 帮我改这段, translate and polish, 翻译并润色, 中译英, 英译中, 缩写, 扩写, condense, expand, 去 AI 味, 去ai味, 降AI, humanize, make it sound less like AI, 逻辑检查, 写摘要, 写引言, 起草 introduction, draft abstract, 写 discussion, 图标题, 表标题, caption, 实验分析, results 分析, 审稿, reviewer 视角, cover letter, 投稿信, rebuttal, 回复审稿意见, response to reviewers, or the user uploads a paper/paragraph and asks to improve the expression, language, readability, or academic quality. It should also trigger even if the user only says "帮我润色一下" or "scipilot 帮我改改文字" or "这段太像 AI 写的". Belongs to the SciPilot Skills research skill family.
Use when the deliverable is spoken words to be performed on camera or into a mic: video and podcast scripts, hooks, and CTAs for short-form (TikTok, Reels, Shorts), YouTube long-form, or podcast episodes. Triggers: "write me a reel about X", "I need a hook for my video", "script a 60-second video", "draft a podcast episode", "turn this idea into a script". For motion/timing/shot specs rather than the words, use motion-picture.
Build or maintain a Codex-native self-improving memory loop using global `AGENTS.md`, a persistent memories directory, and optional nightly refinement automation. Use when Codex needs to adapt OpenClaw-style self-improvement ideas to Codex, set up long-term user/profile memory, create `PROFILE.md` / `ACTIVE.md` / `LEARNINGS.md` / `ERRORS.md` / `FEATURE_REQUESTS.md`, add promotion rules from raw learnings into active guidance, or create a recurring memory-refinement automation.
Build a complete, prioritized SEO strategy for a site: topical authority map, winnable-keyword shortlist, content production pipeline, link-building plan, and technical baseline. Use when the user wants to grow organic traffic, asks 'how do I rank on Google', 'why isn't my site getting traffic', 'build me an SEO plan/strategy', 'what keywords should I target', 'how do I get backlinks', or wants to audit or scale their SEO — even if they never say the word 'strategy'. Produces an SEO Strategy Document with a scored keyword shortlist, a phased action plan sequenced by dependency, and a measurement plan.
Serenity unified framework for supply-chain chokepoint investing v3.0. Distills the best of 10 GitHub repositories, with a nine-step workflow, criteria for good chokepoints, anti-hallucination mechanisms, and adaptation for A-shares, US stocks and global markets. Trigger words: "/serenity", "look at it the Serenity way", "supply chain / chokepoint / bottleneck analysis"
Scaffold a complete knowledge system. Detects platform, conducts conversation, derives configuration, generates everything. Validates against 15 kernel primitives. Triggers on "/setup", "/setup --advanced", "set up my knowledge system", "create my vault".
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.