Skills
97,725 packaged capabilities and playbooks for agents. Install any of them with one command.
All
Generate a consistent multi-angle product image set (front, back, left/right side, top, 45°, detail macro, packaging, in-hand) with identical lighting and background, optionally combined into a grid. Use for 多角度图, 多视角套图, 360 view, angle set, gallery images.
Review code changes through multiple AI models before sending to human reviewers. Use when the user asks to review a branch with sub-agents, run a multi-model code review, get AI review feedback, or mentions "review my branch," "sub-agent review," or "iterate on the code."
Based on the East Money database, it supports natural language screening of A-shares, Hong Kong and US stocks, funds, bonds and other assets, with multi-dimensional indicator filtering covering technical, news, fundamental and market sentiment factors. Suitable for global asset screening, cross-market monitoring, portfolio construction and strategy backtesting. Results include data descriptions and a csv file.
Use when a Materialize object is behind wall-clock time — high lag in mz_wallclock_global_lag, a stale or lagging materialized view, index, or sink, a freshness alert, results that keep showing old data, or a dataflow whose output cannot keep up with its input.
Generate and edit images using Google Gemini (Nano Banana Pro / Nano Banana 2). Automatically selects the best model based on prompt complexity. Supports text-to-image generation, image editing with reference images, configurable aspect ratios, 1K/2K/4K output, Google Search grounding, and batch generation. Works with both Gemini Developer API and Vertex AI.
A methodology for network-layer poisoning and protocol spoofing attacks during internal network penetration testing. Use when network poisoning is needed to obtain NTLM authentication or carry out a man-in-the-middle attack. Covers ARP poisoning, DHCP poisoning, DHCPv6/mitm6 spoofing, DNS spoofing, ADIDNS record injection, WPAD hijacking, WSUS poisoning, ICMP redirection, and NBT Name Overwrite. Usually used as a prerequisite for an NTLM relay attack.
Deprecated alias. This skill was renamed to `deepline-scoring`. Use when a user or script still invokes `/niche-signal-discovery`: immediately load and follow the `deepline-scoring` skill instead. Triggers: ICP analysis, niche signals, won vs lost analysis, differential signals, signal discovery, ICP signal report, account scoring signals, lead scoring, first-party signals, buyer signals.
Build, test, run, and flash NightRun — a bare-metal, no_std Rust UEFI application that boots straight into a local LLM (Llama 3.2, Qwen3, or Granite 4.1) with no operating system underneath. Use when the user wants to build/flash a bootable NightRun USB or Raspberry Pi 5 SD image, convert a GGUF model into the `.nrm` container, run/debug the inference engine on the host or in QEMU/OVMF, or troubleshoot no_std kernel/tokenizer parity issues in the NightRun codebase. Triggers on: "nightrun", "boot into an LLM", "bare-metal LLM runtime", "UEFI LLM appliance", "nrconvert", "nrhost", "cargo xtask", "nrm model file", "flash a bootable LLM USB".
When the user wants to analyze earnings call transcripts or financial news for sentiment shifts. Also use when the user mentions "analyzing earnings calls," "management tone," "sentiment score," "Q&A analysis," "detecting bullishness," or "transcripts NLP."
Write prose clean of AI texture, or audit a draft through two gates: source fidelity (every quoted or attributed line verified against primary sources, four verdicts, correction log) and taste (mechanical tell scan plus a nine-dimension scorecard). Use when the user says "de-slop this", "slop check", "AI smell", "sounds like ChatGPT", "make it human", "humanize this", "sounds robotic", "reads like AI", "kill the em dashes", "verify the quotes", "fidelity check", "run the audit", or "taste pass", and when writing or rewriting X posts, threads, articles, essays, fiction scenes, book chapters, or landing copy that must not read AI-generated. Works at any length; the fidelity gate activates whenever quotes or attributions are present. Do not use for voice imitation (use a style-converter skill), UI design, or proving a text was AI-written: the banned list is house style, not a detector.
(NS) Generate E2E test task markdown for Cypress/TypeScript during SDD planning after frontend tasks exist. Use when planning Cypress coverage per feature — not for writing specs (use ns-code-e2e-tests). Requires data-testid contract from frontend task. Device-aware desktop/tablet/mobile structure.
(NS) Generate backend unit and integration test task markdown after implementation tasks exist. Use during SDD planning for PHPUnit/Laravel or equivalent backend test coverage per feature — not for E2E (use ns-pm-e2e-test-task-generator). Requires requirements and related backend task. Documents cross-tenant and auth P0 scenarios when applicable.
Nx monorepo management skill for AI-native development. This skill should be used when working with Nx workspaces, project graphs, affected detection, code generation, and caching. Use when analyzing dependencies, running affected commands, generating code, configuring Nx Cloud, or optimizing build performance. Invoke nx-mcp tools for documentation queries.
Use NetExec (nxc) / CrackMapExec (cme) for bulk Windows/AD network operations. Use when you need bulk credential validation, password spraying, remote command execution, share enumeration, or credential dumping. nxc is the successor to CrackMapExec and supports SMB/WinRM/LDAP/MSSQL/SSH/RDP/FTP/WMI and other protocols. In domain penetration testing it is used for bulk operations — validating credentials or executing commands across an entire subnet with a single command. Use this skill for scenarios involving CrackMapExec, NetExec, nxc, cme, password spraying, bulk SMB, or domain user enumeration
Job-hunting toolkit. Splits "find a role → sign an offer" into six independent sub-skills: ⓪ Job Hunt for batch role discovery, ① Job Description for decoding a specific JD, ② Resume for a targeted resume, ③ BQ for behavioral interview prep, ④ Offer Compare for comparing multiple offers, ⑤ Salary Negotiation for negotiating the package. All job-hunting requests (finding a job, whether to apply, revising a resume, interview prep, choosing an offer, salary negotiation) enter here and are routed to the matching sub-skill. Keywords: 求职, offer, career, job hunt, LinkedIn jobs, JD, resume, CV, behavioral interview, BQ, STAR, offer compare, salary negotiation, 薪资谈判, 选 offer.
Karpathy-style LLM Wiki with multi-vault sqlite registry and Socratic wizard. Also addressable by the short alias OMW. Manages multiple knowledge vaults (markdown or Obsidian). On invocation, infers user intent from registry state — first-time users get a setup wizard, returning users go straight to operations. Supports memo-mode (lightweight notes) and wiki-mode (Karpathy's raw/wiki/index/log pattern with ingest/query/lint). Trigger phrases — English "open my wiki", "ingest this", "find a note about X", "what does my wiki say about X", "omw", "use omw", "/omw"; Korean "위키 열어줘", "이거 정리해줘", "X 관련 노트 찾아줘", "위키에 물어봐", "오엠더블유", "오엠더블유 켜줘". Also fires when the user pastes long-form content and asks to save it.
TRIGGER THIS when preparing for new hires, planning first-day/week/month activities, creating onboarding schedules, building 30-60-90 day plans, designing welcome programs, organizing training schedules, or setting up buddy/mentorship programs. Designs comprehensive, engaging onboarding experiences that accelerate time-to-productivity and build belonging.
Guide to the OpenAI Agents SDK for JavaScript and TypeScript (@openai/agents, @openai/agents- realtime), covering text agents and realtime voice agents. Covers agents, tools defined with Zod schemas, multi-agent workflows and handoffs, input and output guardrails, structured outputs, streaming, human-in-the-loop patterns, and integrations for Cloudflare Workers, Next.js, React, Hono and Node.js, with templates for each agent type plus error handling and debugging strategies. Prevents 9+ documented errors including Zod schema type errors, MCP tracing failures, agent infinite loops, tool call failures, and schema mismatches. Use when building or debugging agent workflows with this SDK, adding tools or handoffs, implementing realtime voice agents, or hitting those errors.
OpenTelemetry Collector deployment, configuration, and troubleshooting for Coralogix users. Use when writing or debugging collector configs — the `coralogix` exporter (`domain:` vs `endpoint:`, `${env:CORALOGIX_PRIVATE_KEY}` bracket syntax, `coralogix/resource_catalog` variant), the universal processor chain, agent → cluster-collector → gateway topology, spanmetrics/tail_sampling/k8sattributes placement, component stability checks, data-safety/redaction routing including `url_sanitizer`, `sanitize_span_name`, `allow_all_keys`, and before/after validation for broad sanitizer over-sanitization, spanmetrics cardinality protection, and Coralogix-specific presets. Covers the `otel-integration` Helm chart (EKS/GKE/AKS/OpenShift, GKE Autopilot Warden, EKS Fargate), ECS EC2 daemonset, ECS Fargate sidecar, Linux/Windows/macOS standalone, Docker, and the universal installer. Not for OTTL authoring (use the `opentelemetry-ottl` skill) or OpAMP supervisor/Fleet Manager internals beyond the config-precedence callout.
Optimize and compress 3D assets for the web — shrink glTF/GLB models and textures using gltf-transform, gltfpack/meshopt, Draco, and KTX2/Basis, and optionally convert a model to an R3F component with gltfjsx. Use when a model is too heavy, loads slowly, or needs to be web-ready before adding it to a Three.js/R3F scene.
Iteratively optimizes prompts for token efficiency by maximizing information density - reduces verbosity, removes redundancy, tightens phrasing while preserving semantic content. Use when asked to compress, shorten, reduce tokens, tighten, maximize density, increase information density, or make a prompt more concise.
Central entry point of the afonsoft agent harness. Use when starting a new project, resuming an existing one, planning features/Epics/releases, or running any multi-step agent-driven work. Validates and reconciles SPECs (SDD), audits the codebase and harness for gaps (security, architecture, performance, hygiene), proposes improvements, fragments work into GitHub Issues, delegates implementation/QA/review to specialized skills, and re-validates everything until delivery. Also use to review unapproved SPECs, reconcile open GitHub Issues with code, or run a final gap check before closing a release.
Comprehensive OSINT methodology for external red-team operations and authorized attack-surface assessments. Covers the 5-stage recon pipeline (seed discovery, asset expansion, enrichment, exposure analysis, reporting), asset-graph discipline with 29 asset types, severity rubric (CRITICAL/HIGH/MEDIUM/LOW/INFO), confidence upgrade workflows, time budgeting, asset-level triage rules, scale-based tactics, identity-fabric mapping (Entra/Okta/ADFS/Google/SAML/M365 Teams+SharePoint+OAuth), API and auth-map methodology, JavaScript deep analysis, mobile attack surface, cloud attack surface, breach×identity correlation, detectability tagging, detection-aware probing (back-off, persona rotation), read-only validator discipline, WAF/CDN bypass + origin discovery, vulnerability prioritization (CVE/EPSS/KEV), phishing infrastructure planning + pretext development, bug bounty submission templates, client deliverable templates with risk translation, threat-actor investigation (incl. RU/CN pivots), cryptocurrency tracing, image/video forensics, chronolocation. Use when planning or executing reconnaissance against authorized targets, mapping an organization's external attack surface, investigating a person/entity, tracing crypto flows, geolocating media, or performing attribution work.
Find actively maintained open source alternatives to any paid SaaS tool or commercial API. Use whenever a user mentions wanting to replace, swap out, self-host, or find a free/open source version of a paid tool—including tools like Datadog, Algolia, Twilio, Stripe, Auth0, Segment, Mixpanel, Intercom, PagerDuty, LaunchDarkly, SendGrid, Cloudinary, or any other commercial developer tool or API. Also trigger when the user asks things like “Is there an open source version of X?”, “I don’t want to pay for X,” “self-host alternative to X,” or “What can I use instead of X for free?” Runs parallel TinyFish agents to search GitHub and developer communities, then checks each candidate’s real health—last release, contributor count, stars, Docker support, and feature parity—and returns a ranked comparison with a plain English gain/loss summary.
Use PadBuster to perform Padding Oracle attacks. Use when a web application uses CBC-mode encryption and has a Padding Oracle vulnerability. PadBuster can automatically decrypt ciphertext and forge valid ciphertext for arbitrary plaintext, making it suitable for encrypted cookies, tokens, and URL parameters. Use this skill for any scenario involving Padding Oracle attacks, CBC ciphertext decryption, or cookie forgery
Pear Protocol pair/basket trading on Hyperliquid — headless agent-wallet auth (SIWE), V3 gateway REST + Orchard MCP server: markets, pair ratios, positions, orders, and trade execution. Use when the user wants to pair trade (long one asset / short another), check pair ratios, correlation or funding, or manage Pear positions (e.g. long BTC short ETH, check my Pear positions, find trending pairs).
Use this skill whenever the user wants to peer program, pair program, or do collaborative coding where THE HUMAN writes the code and Claude only assists. Trigger on phrases like "pair program", "peer program", "let's code together", "I'll write you review", "drive me through this", "I want to be the driver", "you navigate I type", "teach me by letting me write it", or any workflow where the human is the typist and Claude is the navigator/reviewer. Also trigger when the user wants to learn by writing code themselves with Claude's guidance, or explicitly says Claude should NOT write the implementation. Use this even if the user doesn't say "skill" — the inversion of the usual write/review roles is the signal.
Comprehensive penetration testing workflow using Kali Linux tools via MCP. Guides agents through reconnaissance, vulnerability assessment, exploitation, privilege escalation, and reporting. Use when conducting security assessments, CTF challenges, vulnerability testing, or red team exercises.
Cross-platform Python library for differentiable quantum computing. Integrated with machine learning libraries like PyTorch, TensorFlow, and JAX. Designed for quantum machine learning (QML), variational algorithms, and hardware-agnostic quantum programming. Use for Quantum Neural Networks (QNNs), Variational Quantum Algorithms (VQE, QAOA), hybrid classical-quantum machine learning, quantum chemistry calculations, benchmarking quantum algorithms, optimizing quantum control pulses, and investigating QML phenomena like Barren Plateaus.
This skill should be used when working on any design, marketing, copywriting, writing, or communication task — including: "run PFD", "apply PFD", "use Perception-First Design", "do a PFD analysis", "run the derivation protocol"; when designing interfaces, landing pages, emails, ads, or visual systems; when writing or reviewing marketing copy, headlines, CTAs, proposals, pitches, or presentations; when evaluating conversion problems, bounce rates, engagement, or persuasion; when doing brand work, content strategy, or any client communication; or whenever cognitive load, first impressions, processing fluency, trust, or decision architecture should inform the work.
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.