Skills
97,501 packaged capabilities and playbooks for agents. Install any of them with one command.
All
【Remove AI traces】Detect AI-like writing by scanning copy sentence by sentence for machine traces across 20 items, including stance, structure, wording, content, and mass-message-like private-domain tone. Report findings in the original order; diagnose only by default. If the user asks for edits, ask what they intend to achieve with each sentence, then rewrite it in their own words. Covers short-video scripts, Moments posts, direct messages, Xiaohongshu posts, and long-form WeChat Official Account articles. Trigger when the user says “does this sound like AI,” “does this read like a machine wrote it,” “does this sound fake,” “does this Moments post look like a mass message,” or “remove the AI feel.”｜小爷出品
【Idea desk】Manage the stage where “an idea grows into a usable content topic”: jot down spontaneous ideas, help the user clarify their intuition through conversation, write an expanded draft using a fixed structure, assess who the topic is for, why it could attract traffic, and why it could monetize, then route it to short video, Moments, or private-domain scripts. Archive the master version once the final draft is complete. It does not handle writing drafts for topics that are already decided. Use when the user says “note an idea,” “develop this idea,” “organize this idea,” “is this topic worth doing,” or “idea desk.”｜小爷出品
【Personal brand positioning】Diagnose IP positioning using seven verifiable criteria: business anchor, authentic core, niche depth, emotional connection, stage of audience mindshare, memorability, and psychological account. Assess whether a persona holds up and can withstand the close-range trust test of private-domain business. In consultation mode, review the user’s own account; trigger when they say “review this persona positioning,” “how should I position my IP,” “does this persona hold up,” “is my account going in the right direction,” “should I pivot,” or “analyze my target audience.” In operator mode, help develop positioning for others; trigger when they say “position this IP,” “I got a new project; help me with preliminary positioning,” or clearly describe a third party’s business.｜小爷出品
【Execution diagnosis】When the user knows what to do (post the first video, add the first group of people, sell the first product, or start the first livestream) but still does not act, identify which of six “not getting started” patterns is holding them back and give the smallest starting action for that pattern—not empty pep talks. Use when the user says “I know what to do but don’t do it,” “I’m always preparing,” “I want to change direction again,” “I’ll start after this course,” “let me polish it a bit more,” or “my circumstances don’t allow it right now.”｜小爷出品
【Bridging】Link a skill or the entire skills/ directory via symlinks to the locally installed Agent hosts (the shared entry point ~/.agents/skills and dedicated entry points such as Claude Code, WorkBuddy, and Trae; generate a thin bridge for Grok). Also check bridge status and remove bridges. Use when the user says “bridge this skill,” “connect the skill to Codex/豆包/WorkBuddy,” “make it available to all Agents,” “is the bridge ready?” or “remove the bridge.”｜小爷出品
【Model review】Review team and commission-based model design across five dimensions: model classification, compliance red-line self-check, top-level and revenue-sharing structure, C-end health, and screening for harmful models. Assess whether an agency, distribution, or referral system is viable and expose traps such as “passive income,” “low-to-high conversion,” and “handing down customer resources.” For compliance, provide verification paths only; do not draw conclusions. Trigger when the user says “review this commission model,” “should I use an agency model,” “is this team policy compliant,” “how should we split revenue,” “is this a pyramid scheme,” or “I want to design a distribution system.”｜小爷出品
【WeChat Official Account formatting】Convert Markdown manuscripts into HTML that can be pasted directly into the WeChat Official Account backend without breaking the layout. Includes 15 formatting styles; generate 1, 6 recommended, or all 15, with an overview page. Trigger when the user says “convert this to a WeChat Official Account version,” “format this for an Official Account,” “generate Official Account HTML,” “how should I format this article for an Official Account,” or “WeChat formatting.” Formatting only; do not change the content.｜小爷出品
【Getting started in self-media】A starter SOP for newcomers to private-domain business and short video. Use verifiable principles—prioritize early wins, start with familiar-audience traffic, treat the first 100–200 customers as practice costs, and design a transition with a newcomer offer—to give complete beginners or those just starting out with no results a practical sequence of first steps. Head off self-congratulatory preparation such as “build the whole system first” or “learn everything before taking action.” Trigger when the user says “I’m a beginner; how do I start,” “what’s the first step when starting out,” “beginner starter SOP,” “how should I start my account,” “where do I find my first customers,” or “how do I post on Moments without sounding like I’m selling?”｜小爷出品
【Opening diagnosis】Diagnose and generate short-video openings. First apply a gate to determine whether the content merits a strong opening; then assess the existing opening across six dimensions; finally, use three methods—mining the draft, adding facts, and withholding the answer—to produce 10–15 candidates and recommend the Top 3. Trigger when the user says “how should I write the opening,” “viewers drop off in the first 3 seconds,” “the 5-second drop-off rate is too high,” “what should the first line be,” “revise the opening,” or “the opening is too flat.”｜小爷出品
[Team operations review] Review from an operator’s perspective across five dimensions: fit with the business launch stage, organizational size thresholds, productivity red lines, execution of profit sharing and team morale, and frontline validation. Assess whether team building, productivity design, and launch pace match the actual business stage, filtering out false signals such as “a bigger team is a stronger team,” “good-looking GMV means profit,” and “relationships are enough to manage a team well.” Trigger when users say, “Check whether my team is set up right,” “How should I set commissions?” “Should I expand the team?” “Is the launch pace too fast?” “Is there a productivity problem?” or “Why can’t the team retain people?”｜小爷出品
[Benchmark selection] Find benchmark accounts or businesses genuinely worth emulating. First apply three hard criteria: profitability, scale fit, and account health. Then check that what’s being copied is the business structure (people, products, and channels; pricing; customer acquisition and conversion; delivery and repeat purchases), not surface elements like cover images, titles, and copy. Finally, break it down into an actionable checklist; don’t use “it doesn’t suit me” as an excuse. Trigger when users ask to find benchmarks, who to learn from, whether an account is worth copying, how to imitate a benchmark, or to analyze an account’s business.｜小爷出品
[Playbook by niche] Develop a playbook for a niche: input the industry, average order value, and current situation; get the mindset stage from which revenue comes, the mix of three content functions, product tiers, how traffic moves into private channels, where conversions happen, compliance red lines, and reusable account structures. Includes 8 niche templates (weight loss/weight management, traditional Chinese medicine wellness essential oils, fashion and fast-moving consumer goods, direct-sales conversion and agency-system restructuring for social commerce, small-group knowledge courses, direct sales through private channels in the wellness sector, high-decision-cost non-standard services such as study abroad, and buy-use-separated education such as English for beginners). Other niches use general analysis and mark assumptions with 🔴. Trigger when users ask how to acquire customers in their niche, how to approach an industry, what content to make for their average order value, how to approach weight loss/wellness/fashion/social commerce-to-direct-sales/knowledge courses/study abroad, or to work out a niche playbook.｜小爷出品
[Business case comparison] Historical structural analysis: condense the user’s current business challenge into a “structural fingerprint,” then search business history, industry history, and the XY case library for structurally similar successes, failures, and counterexamples. Compare recurring mechanisms, conditions for success, and limits where they fail to determine whether there is a “standard answer.” Trigger when users ask whether anyone has faced this before, who encountered it in history, whether there is a classic solution or standard answer, or to find an analogy.｜小爷出品
[Private-channel operations] Diagnose day-to-day private-channel operations: content mix and cadence for Moments, account presentation and persona, tag segmentation and tagging, reactivating dormant contacts (proactively and passively), community operations, trust-building through ongoing support, direct-sales vs. social-commerce model assessment, and a seven-element self-check for professionalizing private channels. Review whether your private channels are being “managed” or just “filled with people.” Trigger when users ask what or how often to post in Moments, how to tag contacts, why added friends don’t engage, how to reactivate old customers, why a community is silent, whether their model is direct sales or social commerce, or whether Moments ads could reduce reach or get an account banned.｜小爷出品
[Pre-publication risk check] Review copy, spoken scripts, titles, subtitles, comments, account details, or visual descriptions sentence by sentence before publication. Separately identify what platform automated moderation may flag and whether the content itself has substantive issues. For each issue, provide its location and the smallest corrective change; don’t rewrite the full text or teach ways to evade moderation. Use when users ask whether a post can be published, for a risk check, about reach limits or prohibited terms, for a pre-publication review, or whether a sentence is problematic. Reach-limit and account-ban issues in Moments/WeChat belong to xy-private-ops.｜小爷出品
[Problem brief] Turn vague questions such as “Why is nobody making inquiries?”, “The customer added me but isn’t replying,” and “Can AI automate this?” into a problem brief with a defined subject, conflict, constraints, and feedback channel, then assess what level of the task an Agent could take over. Use when users ask to clarify a problem, whether it can be automated, say they don’t know what’s wrong, ask why nobody is doing something, or ask whether AI can do it for them.｜小爷出品
[Viral content recreation] Break down a benchmark viral video into its structure and transferable formula (rank evidence; don’t make things up), then adapt it with your product’s real selling points to produce a recreation script or AI video-generation prompt. Until usable selling points are provided, deliver only structural analysis—never invent selling points or product appearance. Trigger when users ask to break down a viral video, copy its structure, make one for their product based on it, recreate it with a different product, or write a video-generation prompt.｜小爷出品
[Pick up where we left off] Bring up the saved diagnosis from last time so the user can continue from the previous conclusions. Use with xy-archive. Use when users say “continue from last time,” “where did we leave off?”, “what were the previous conclusions?”, “pick up where we left off,” or “bring up the archive.” Trigger: /xy-resume, “pick up where we left off”, “what did we conclude last time”｜小爷出品
[Multi-perspective roundtable] A focused chat room. Introduce a topic; the moderator identifies missing judgment variables, selects 3–5 theoretical frameworks and their representative thinkers to discuss it together, then adjudicates disagreements. “小爷” (9 principles + atomic retrieval) is always present, and users can call on the “classic four.” There is also an “XY deep analysis” mode that produces a report using four lenses: people, money, structure, and red lines. Use when users ask for help thinking, different perspectives, a chat room, deep analysis, or analysis by four thinkers.｜小爷出品
[Script flow check] Check the flow of short-video transcripts and identify the exact moment viewers may lose track: whether transitions between sections are broken, sections contain filler, or sentences are hard to say aloud. Rate issues 🔴🟡🟢 and provide their locations and fixes. After the report, proactively ask whether the user wants an “annotated revision” (preserve the original, mark edits in pairs, and allow them to be reverted). Trigger when users ask whether a script flows, where viewers might swipe away, whether the middle is wordy, whether the logic breaks, to smooth out a script, or say the spoken delivery feels awkward.｜小爷出品
[Product selection diagnosis] Assess whether a product category is viable and whether the downstream operations can support it. Use a two-axis classification—whether the market needs educating and who holds pricing power—along with supply-chain and total-cost accounting and a three-round paid market test. Filter out false signals such as “I feel it’ll sell” and “the GMV looks good.” Trigger when users ask whether they should pursue a product, to assess a category, how to choose products, whether to calculate supply-chain costs, whether to test the market, or whether poor sales are a product-selection problem.｜小爷出品
[Skill security audit] Scan installed skills in local Agent host directories for signs of overreach: covert traffic diversion, undisclosed commercial relationships, hijacking user instructions, reading sensitive data, or suspicious external calls. By default, generate a report only; isolate a problematic skill in a local quarantine area (recoverable, not deleted) only after the user explicitly confirms. Use when users ask to scan installed skills, check for ads, assess a skill’s safety, clean up problematic skills, or audit their skills.｜小爷出品
[Diagnosing where to slow down] Slow is fast. Help people working in private channels, product selection, or team management distinguish which parts of a task are worth doing slowly themselves (the first 100 customers, the first product test, the first 30 days of Moments posts, doing a role before hiring for it) and which should be done quickly, outsourced, or handled with tools. Use when users ask for a slower but steadier method, whether they’re taking shortcuts, whether something can be outsourced, templated, or bulk-sent, why others can do it faster, or say they want to save effort.｜小爷出品
[Sync updates] Sync the XY workbench: rebuild atomic exports and re-bridge all repository skills/ to each host so updated or newly written skills are immediately available in Claude Code, Codex, WorkBuddy, Grok, and other endpoints. Once the repository is connected to GitHub, also perform two steps: safe push (prevent internal files from being pushed when a parent directory is initialized as a repository) and pull updates. Use when users ask to update XY, sync skills, sync newly written skills across endpoints, push to GitHub, pull the latest, or check for a new version.｜小爷出品
[Concept breakdown] Break down overused terms in private-channel, short-video, and IP circles (private channels, direct-sales conversion, social-commerce conversion, referral promoters, order-based followers, traffic-driving/profit products, productivity, the lonely period, the experience of winning, mindshare, spiritual core, passive income, and targeted traffic) until they can be expressed in actions and numbers. Identify which are old ideas in new packaging and which questions have no answer in words alone. Use when users ask what private channels really are, what a term means, to break down a concept, whether someone means the same thing they do, or whether something is actually new.｜小爷出品
[Traffic diversion diagnosis] Diagnose how to design a path from public traffic to private channels. Assess five areas: differences in platform risk controls, diversion intensity by account stage, risk levels of tactics, traffic quality filtering, and downstream handling. Filter out naive plans such as copying competitors’ diversion scripts, assuming more traffic is always better, or collecting contact details first and figuring out the rest later. Also handles short-video account launch and cold start (start here if a new account has no traffic or has been tagged incorrectly). Trigger when users ask why a new account has no traffic, how to launch an account, whether a diversion method is right, how to design a traffic path, how to direct traffic to WeChat, whether a method is risky, how intense diversion should be, or how to move from public traffic to private channels.｜小爷出品
[Knowledge base] Turn a user’s local folders (customer records, product lists, Moments content, course transcripts, historical versions of agency policies, livestream reviews, etc.) into a knowledge base that AI can reliably search, organize, and version-check. If there are no materials, create a minimal structure; if materials exist, generate a “knowledge base guide” that only points the way. Then manage new materials, find answers, identify the latest versions, audit, and streamline. Trigger when users ask to build a knowledge base, have AI understand their files, add something to the knowledge base, search it, identify the latest version, update it, or shorten the guide. Trigger phrases: /xy-vault, /知识库.｜小爷出品
[Multi-platform Agent workbench setup] Organize a project into a consistent, maintainable Agent workbench across Claude Code / Codex / Grok / generic Agents (~/.agents/skills): review rules files (CLAUDE.md / AGENTS.md), identify the source of truth for skills, standardize naming, generate bridges, and verify each item. Use when users want to migrate to Codex or Claude Code, make it work with Grok, standardize AGENTS.md, tidy up a messy Agent workbench or scattered skills, or connect Claude, Codex, and Doubao.｜小爷出品
Use when working with Turso yield injection or synthetic failure injection machinery for safe resumable state-machine boundaries, including YieldInjector, FailureInjector, YieldPointMarker, inject_transition_yield!, inject_io_yield!, inject_transition_failure!, CommitYieldPoint, CheckpointYieldPoint, CursorYieldPoint, deterministic interleaving tests, abandoned statement/commit tests, and concurrent-simulator yield plans.
Replace with a trigger-style description of when this skill should activate. Be specific — this is what the agent uses to decide whether to load the skill. Example: "Sui TypeScript SDK integration. Use when writing, reviewing, or debugging TypeScript code that interacts with Sui RPCs, transactions, or on-chain state."
Researches products, prices, brands, and categories on Zalando (the European fashion marketplace) using the Crawlora API, returning clean JSON. Use when the user asks to find a product, browse a category or brand, autocomplete a search, or resolve a Zalando storefront market—instead of scraping Zalando pages.
Use when scaling a dedicated Kubernetes cluster (Linode LKE or AWS EKS) on Zeabur — listing node pools, changing node counts, or adding/removing node pools. Use when user says "scale my cluster", "add nodes", "remove nodes", "resize node pool", "加節點", "擴容", "縮容". Do NOT use for renting new servers (use zeabur-server-rent) or for single-VM dedicated servers (they have no node pools).
Use only when a focused Sales workflow has selected a present and connected ZoomInfo connector, or the user explicitly asks for ZoomInfo company search, contact search, enrichment, intent, similarity, recommendations, or research. Do not use when another enrichment source is selected.
Reverse-engineer one or more local reference videos into verifiable, generatable source footage, shot frame boards, an asset library, and a segmented generation plan that can be reviewed directly in an Agent conversation. Use when the user says 'reverse this video', 'reverse these videos', 'how do I replicate this video', 'break the video into prompts', or 'reverse this video', or needs shot-by-shot analysis, evidence extraction, or audit of the reverse-engineering package before replicating a video; not for directly generating videos, publishing assets, general editing, or burning in subtitles.
Generates publication-quality academic figures (framework diagrams, pipeline illustrations, system architectures, method overviews) from a paper's method text and a target caption, using a local PaperBanana multi-agent pipeline (Retriever → Planner → Stylist → Visualizer → Critic).
Multi-perspective academic paper review with dynamic reviewer personas. Simulates 5 independent reviewers (EIC + 3 peer reviewers + Devil's Advocate) with field-specific expertise. Supports full review, re-review (verification), quick assessment, methodology focus, Socratic guided, and calibration modes. Triggers on: review paper, peer review, manuscript review, referee report, review my paper, critique paper, simulate review, editorial review, calibrate reviewer, reviewer calibration, measure reviewer accuracy.
Simulates academic peer review, evaluating papers across Originality, Methodology, Results, and Writing to provide Major/Minor Revision recommendations with actionable feedback. Triggers when a user asks to "review my paper," "simulate peer review," or "give my paper a peer review."
Orchestrator for the full academic research pipeline: research -> write -> integrity check -> review -> revise -> re-review -> re-revise -> final integrity check -> finalize. Coordinates deep-research, academic-paper, and academic-paper-reviewer into a seamless 10-stage workflow with mandatory integrity verification, two-stage peer review, and reproducible quality gates. Triggers on: academic pipeline, research to paper, full paper workflow, paper pipeline, end-to-end paper, research-to-publication, complete paper workflow.
Find, screen, and synthesize academic papers, scientific literature, technical reports, preprints, clinical or biomedical publications, and evidence around a research question. Use when the user asks to find papers, summarize literature, compare methods, identify seminal or recent work, extract claims from studies, or build a source-grounded academic/scientific evidence brief.
Perform deep WCAG 2.1 AA compliance audits on UI code. Use this skill when the user wants an accessibility review, WCAG audit, a11y check, screen reader compatibility review, or keyboard navigation assessment. Also trigger when the user mentions 'accessibility', 'a11y', 'WCAG', 'screen reader', 'keyboard navigation', 'color contrast', 'ARIA', or asks 'is this accessible'. This goes beyond the design reviewer's basic accessibility checks to provide comprehensive WCAG criterion-by-criterion analysis
Design and audit accessibility conformance programs for apps, SaaS, websites, desktop software, games, marketplaces, and developer tools. Covers WCAG targets, VPAT/ACR, audits, assistive-tech testing, design systems, issue severity, ownership, remediation, release gates, procurement evidence, and customer communication. Use when accessibility needs durable governance beyond one-off audits.
Accessibility testing toolkit using Selenium WebDriver 4+ with Java 21+ and axe-core engine. Use when asked to validate WCAG 2.2 AA compliance, scan pages or components for a11y violations, test keyboard navigation, audit color contrast, check ARIA semantics, generate accessibility reports, filter axe rules, debug screen reader issues, or implement POUR principles (perceivable, operable, understandable, robust).
Check the Sonilo account's available services, rate limits, free-trial allowance, and usage/billing history. Use before calling a paid Sonilo tool to confirm it's available and whether free-trial runs remain, or when the user asks about their Sonilo usage, limits, or billing
Research ONE target company as an outbound prospect and produce a decision-ready account dossier: signal stack, fit against YOUR ICP, decision-makers, lead score, risks, next actions and first-touch drafts in the prospect's language. Use when asked "research this account", "build a dossier on [company]", "is [company] a good fit for us", "who do we contact at [company]", "pre-call research on [company]", "should we go after this account", "account brief for outbound", or when handed a company URL/name to qualify. This is PROSPECT-side research (them, as a buyer). For vendor-side research on why your own customers buy from you, use `customer-intelligence` instead.
Research a company or person and get actionable sales intel. Works standalone with web search, supercharged when you connect enrichment tools or your CRM. Trigger with "research [company]", "look up [person]", "intel on [prospect]", "who is [name] at [company]", or "tell me about [company]".
Use when you need to actually run security tests against a running application you own or are authorized to test — the shared execution harness for lab/proxy/scanner setup, safe non-destructive execution, and standardized proof-of-concept capture that the find/exploit specialist skills rely on.
How to use the Adaptyv Bio Foundry API and Python SDK for protein experiment design, submission, and results retrieval. Use this skill whenever the user mentions Adaptyv, Foundry API, protein binding assays, protein screening experiments, BLI/SPR assays, thermostability assays, or wants to submit protein sequences for experimental characterization. Also trigger when code imports `adaptyv`, `adaptyv_sdk`, or `FoundryClient`, or references `foundry-api-public.adaptyvbio.com`
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.