Skills
100,696 packaged capabilities and playbooks for agents. Install any of them with one command.
All
Run a 1:N face search with Didit's standalone Face Search API (POST /v3/face-search/) against every face enrolled in the application: verification sessions, saved standalone calls, faces uploaded to User profiles, face lists, and retained biometric templates. Use to detect duplicate accounts, check whether a face already exists, prevent repeat registrations, screen a face against the blocklist, or investigate fraud rings. Returns up to 5 ranked matches with similarity percentages, match source and session details, plus blocklist and duplicate-face warnings.
Use for Ding-style (钉内/钉外) workplace reminders rooted in the 7.5万字 essay 《置身钉内》 and VP response 《置身钉外》. Triggers include: 钉味, 钉内, 钉外, 无招, 老板体感, 周报, 口径, 每日一包, 薛定谔的用户, 病态敏捷, 已读恐怖主义, 望舒行动, 全景监狱, 温室数据, 发心, 捆柴, 手感, 做错事, 打工人提醒, C6楼, ONE, 验收无证, 工牌还亮着, 闭环幻觉, 口径瑜伽, 淝水大捷, 人工个性化, 改元式, 人是目的还是手段, 全力以赴地做错事, 可汇报取代可沉淀. Do not use for pure coding tasks with no workplace/org/delivery context.
Automate all Feishu (Lark) channels. Use the lark-mcp tool to work with Feishu Bitable, send messages, manage documents, create groups, automate workflows, and more. Use this skill when the user needs to operate the Feishu platform, sync data to Feishu tables, send Feishu notifications, manage Feishu documents, or automate Feishu business processes.
Implement Git branching strategies, PR workflows, and release management patterns. Configure GitFlow, trunk-based development, or GitHub Flow for team collaboration. Use when establishing version control workflows or improving development team collaboration.
A specialized skill for reducing AIGC detection rates in Chinese academic writing (papers, journal articles, theses, etc.). It distills patterns from real paper-rewriting experiments (reducing the AIGC rate from >50% to 11%) and draws on the systematic pattern classification in Wikipedia’s “Signs of AI writing” project. Detects and fixes common patterns in Chinese AI writing, including theory-based openings, formulaic endings such as “this case confirms that,” neatly parallel sentences, templated problem statements, passive-analysis clichés, overly symmetrical structures, redundant summary sentences, vague attribution, filler phrases, generic positive conclusions, and overused AI buzzwords. Use when the user asks to: lower AI detection rates, polish writing by hand, reduce AIGC, humanize text, remove the AI feel, polish academic writing, reduce plagiarism similarity in a paper, check AIGC detection, remove AI traces, lower a paper’s AI rate, remove signs of AI writing, or make academic text more human.
Adversarial review (对抗 review) of a finished deliverable — a change ready for a PR, an investigation report, a recommendation, a batch of record edits — before it is reported to the user. Spawns fresh reviewer subagents briefed to disprove it from three directions (open, over-engineering, residue), verifies and fixes what holds, then reports. Run once per deliverable when it is considered done, and on request at level strong (强力) or max (极致). Not for answers to questions, lookups, or steps inside a task.
Guides use of the icp command-line tool for building and deploying Internet Computer applications. Covers project configuration (icp.yaml), recipes, environments, canister lifecycle, identity management, and bundling a project into a self-contained .icp package (icp project bundle). Use when building, deploying, or managing any IC project. Use when the user mentions icp, dfx, canister deployment, local network, project setup, or bundling/packaging an app as an .icp file. Do NOT use for canister-level programming patterns like access control, inter-canister calls, or stable memory — use domain-specific skills instead.
Explore PostHog's Inbox and act on what it surfaces — the place where signal reports cluster into actionable issues and trends. Use when the user asks "what's in my inbox?", "what should I look at?", "which reports are actionable?", "what's PostHog flagged recently?", asks about a specific report by ID or title, wants to act on / fix / implement a report (turn it into a PR), wants to resolve, dismiss, or snooze a report, or wants to see which signal sources are configured. Also use before you investigate an error, bug, metric change, or support ticket in code, to find research the Inbox already did. Covers listing, filtering, drilling into, and acting on reports, plus pointers to the deeper `signals` skill when raw signals or semantic search are needed. For a personal, ranked top N of reports, use `prioritizing-my-inbox`, which starts from the stored personal ranking (`view=for_you`, `sort=personal`).
Design oRPC v2 APIs contract-first, defining the API shape with oc from `@orpc/contract`, implementing it with implement from `@orpc/server`, and consuming the contract from typesafe clients. Use when a project depends on `@orpc/contract`, when defining a contract with oc, implementing a contract with implement, sharing an API contract between server and client packages, generating a contract from an existing OpenAPI spec, or publishing a typed API client to npm. Biases toward retrieval from the oRPC docs over pre-trained knowledge. For core builder, serving, and client work without a contract, use the orpc skill; for REST/OpenAPI exposure, spec generation, and OpenAPILink details, use the orpc-openapi skill.
Find candidate stocks with a sufficient pullback, an intact trend, and a reasonably sound support structure, then provide observation zones, reversal signals, and invalidation conditions. Suitable for finding buy-the-dip opportunities, re-entering strong stocks, and monitoring pullbacks in trending stocks.
Use when doing any Autodesk Maya task through Python or a bridge, headless or live: maya.cmds or OpenMaya 2.0 scripts, mayapy or maya.standalone batch jobs, a commandPort or MCP connection into a running Maya, "Maya 2027", validating or exporting scenes, or a maya.cmds script that fails after the 2025 to 2027 changes (PySide6, OpenPBR, renamed math nodes, removed flags). Also when a Maya model, rig, shot or render must be judged like a professional before it is called done, or a brief must be routed to the right Maya specialist.
Use when a ZBrush sculpt, scan or AI-generated mesh must become production topology and files (ZRemesher and guides, the 2026 Retopo brush, reprojecting detail with Project All, rebuilding subdivision levels, UV Master or native unwrap, normal and displacement maps with Multi Map Exporter, GoZ, OBJ or FBX export, scale and axes, Decimation Master, "clean up an AI mesh in ZBrush"), or when detail is lost after ZRemesher, a map is upside down or an export arrives at the wrong size.
Use when asked to "shape up", "run a shaping session", "set an appetite", "scope a project without estimates", "betting table", or "ship in fixed cycles". Helps teams escape estimate-driven development and Scrum fatigue. The Shape Up method (created by Ryan Singer at Basecamp/37signals) uses fixed time boxes, variable scope, and collaborative shaping to ship meaningful work predictably.
Use when the user asks about revenue, payments, subscriptions, billing, CRM deals, support tickets, ad spend, production database tables, bucket files, or local file uploads. Also use when the user wants to join PostHog events with external data, or when a query fails because an expected table is missing. The data warehouse can import from SaaS tools, ad platforms, databases, object storage, and uploaded files. Covers checking existing sources, identifying the right source type, and guiding setup.
Dialogue writing and revision for screen and stage (对白/台词/潜台词) — merged from Robert McKee's Dialogue (dialogue as action, said/unsaid/unsayable, exposition as ammunition, beat = action/reaction gerunds, flaws of credibility/language/content/design, character-specific vocabulary and grammar, seven scene analyses), Richard Walter (economy, subtext, argument, no small talk, no parentheticals, silence), Neill Hicks (listening, on-the-nose exceptions, energy, five-step exercise), Lu Jun (语言要美; 典雅美/通俗美, 对话三要求 动作性/性格化/潜台词), Lajos Egri (dialogue born of three-dimensional character, escalation), Blake Snyder (cover-the-names test), and screwball-comedy dialogue conventions. Use when writing or polishing dialogue, fixing on-the-nose or expository lines, giving each character a distinct voice, converting a two-hander into a trialogue, writing comedy exchanges, or cutting speeches. Chinese-opera lyrics (戏曲唱词) are in sw-chinese-opera-banqiang, not here.
Implement Syncfusion Blazor Accumulation Chart (SfAccumulationChart) for proportional data visualization. Use this when creating pie charts, doughnut charts, funnel charts, or pyramid charts. This skill covers data labels, legends, tooltips, slice grouping, and part-to-whole visualizations in Blazor applications.
Gene-set enrichment analysis — GO (Biological Process, Molecular Function, Cellular Component), KEGG, Reactome pathway enrichment via clusterProfiler, gseapy, ORA, GSEA, plus g:Profiler for enrichment/ID conversion/ortholog mapping/SNP-to-gene annotation. Use for interpreting DEG lists, screen hit lists, or any gene-list-to-pathways query. Includes simplify-cutoff handling and union-vs-total denominator conventions for percent-DE questions.
Access comprehensive LaTeX templates, formatting requirements, and submission guidelines for major scientific publication venues (Nature, Science, PLOS, IEEE, ACM), academic conferences (NeurIPS, ICML, CVPR, CHI), research posters, and grant proposals (NSF, NIH, DOE, DARPA). This skill should be used when preparing manuscripts for journal submission, conference papers, research posters, or grant proposals and need venue-specific formatting requirements and templates.
Audits the full lifecycle of bearer/API tokens protecting an internet-exposed HTTP API — minting entropy and format, hash-only server storage, transport rules, scoping, expiry and rotation, revocation, per-key rate limits, and log hygiene — with read-only evidence commands and hardened reference implementations.
Curate a Chinese reading digest from a fixed bundle of RSS and Atom feeds, with a strong preference for AI agent thinking, frontier AI commentary, deep interviews, and non-boring high-signal essays. Use when Codex needs to pull the latest week's posts by default, or a specific day's posts when explicitly requested, summarize them, score each article on a 10-point scale, and output only the posts scoring above 7 in a concise Chinese daily-brief style.
Manages MongoDB Atlas Stream Processing (ASP) workflows. Handles workspace provisioning, data source/sink connections, processor lifecycle operations, debugging diagnostics, and tier sizing. Supports Kafka, Atlas clusters, S3, HTTPS, and Lambda integrations for streaming data workloads and event processing. NOT for general MongoDB queries or Atlas cluster management. Requires MongoDB MCP Server with Atlas API credentials.
Improve the quality and type safety of TypeScript/JavaScript being written or refactored—parse, don’t validate; make illegal states unrepresentable; use exhaustive matching; separate pure functions from IO; and maintain public API hygiene. Use when asked to refactor, clean up, improve, harden, or review the design or quality of code or an API. Not for hunting bugs (use /code-review) or for vigiles’s own spec/lint rules.
Analyzes code diffs and files to identify bugs, security vulnerabilities (SQL injection, XSS, insecure deserialization), code smells, N+1 queries, naming issues, and architectural concerns, then produces a structured review report with prioritized, actionable feedback. Use when reviewing pull requests, conducting code quality audits, identifying refactoring opportunities, or checking for security issues. Invoke for PR reviews, code quality checks, refactoring suggestions, review code, code quality. Complements specialized skills (security-reviewer, test-master) by providing broad-scope review across correctness, performance, maintainability, and test coverage in a single pass.
Create, read, update, move, and manage kanban board feature files backed by markdown with YAML frontmatter. Use when working with kanban boards, task/feature tracking, `.devtool/features/` directories, feature files with status/priority frontmatter, or any project management tasks involving markdown-based kanban workflows.
Linux kernel architecture skill for subsystem overview. Use when navigating kernel source, understanding boot flow, initcalls, or major subsystems (VFS, scheduler, MM). Activates on queries about kernel architecture, boot process, initcall levels, subsystem layout, or where code lives in linux.git.
Build comprehensive randomization lists for creative entropy. Use when you need to create or expand lists of story elements (professions, locations, objects, names, etc.) for use with entropy tools. Leverages research sources like Kiwix/Wikipedia to build lists with good variety and size.
Route BigQuery-backed Looker Studio work into one stakeholder-reporting packet: dashboard-spec, slow-dashboard triage, refresh-shape choice, audience split, or exec-handoff. Use when the user needs KPI boards, PM/ops reviews, marketing / GTM reporting, product funnel summaries, or game/live-ops telemetry dashboards on top of curated BigQuery data. Route KPI explanation to `data-analysis`, repeated anomaly hunting to `pattern-detection`, telemetry/alerting coverage to `monitoring-observability`, and semantic-platform choice to `survey`.
Write scripts for AI short dramas: turn novel-outline episode summaries into structured scenes and beat sequences, alternating action beats and dialogue lines, with every line tagged by speaker and tone. Episode durations are calculated deterministically from speaking speed. Produces script.json, Markdown, and a one-page review report (duration dashboard, episode scripts, scene overview, and dialogue grouped by character, with JSON export). The script manages the drama; storyboards manage the shots—this skill does not create storyboards, shot numbers, or generation prompts; those belong to the next-layer storyboard skill. All 10 quality gates are checked deterministically by scripts (episode duration ±15%, single lines ≤35 characters, valid speakers, hooks written down, concrete hook payoff within the first 3 beats (hookBeat claims the cold open), at least one action beat per scene, payoff claims, and reconciliation of characters/scenes/lighting/props with upstream data…). If outline.json exists, use seed to prefill each episode's skeleton; with --art, reconcile lighting states too. Zero dependencies, zero API keys; uses the current session quota. Use when asked to write a script, dialogue, scenes, or screenwriting for an AI short drama.
Odoo 17 development reference for Python models and ORM (search, domain, read_group, compute fields), XML/CSV data and views, OWL/JS client code, QWeb reports, security (ACL, record rules, groups), cron and server actions, migrations and module upgrades, tests, i18n, and performance. Use this skill whenever work involves Odoo 17 or custom addons—even if the user only pastes a traceback, mentions addons/ or __manifest__.py, describes form/tree/kanban/XML errors, HTTP controllers, or business rules on models—including building features, fixing bugs, refactoring, or reviewing addon code. Includes CSS/SCSS asset authoring and review for Odoo addons.
Use when asked to "set OKRs", "objectives and key results", "quarterly OKR planning", "align objectives", "measure OKR progress", or "focus priorities with OKRs". Helps teams focus on what matters most and create a cadence of progress. The OKR framework (originated by Andy Grove at Intel, popularized by John Doerr at Google) creates alignment, focus, and learning cycles. Christina Wodtke's Radical Focus approach emphasizes simplicity and avoiding common pitfalls.
Load project context, list existing specs and changes, and search for capabilities and requirements. Use when users ask about project status, existing specs, ongoing changes, available capabilities, or need to discover context. Triggers include “openspec context,” “what specs are there,” “show changes,” “list capabilities,” “project context,” “find specs,” “what's in the specs,” and “show specs.”
Expert in building Retrieval-Augmented Generation systems. Masters embedding models, vector databases, chunking strategies, and retrieval optimization for LLM applications. Use when: building RAG, vector search, embeddings, semantic search, document retrieval.
Use when a Godot 4 game goes online or co-op: host and join with ENet, WebSocket for a web build, RPCs (@rpc, rpc_id, any_peer), MultiplayerSpawner and MultiplayerSynchronizer, player authority, lobbies and login, a dedicated or headless server export, lag, rubber-banding, prediction or netfox, cheating, bandwidth, Steam, Nakama or WebRTC. Symptoms: players do not appear on the client, 'rpc node checksum failed', 'node not found' on an RPC, desync, the server build is huge, a client never connects.
Use when modeling hard surface in ZBrush 2026 (sci-fi helmet, weapon, prop, armor plates, mechanical parts, kitbash, vents, panel lines, bolts, concept to game asset), when using ZModeler, Dynamic Subdivision and creasing, Live Boolean, Knife, Clip, Slice or Trim curves, IMM, ArrayMesh, NanoMesh or Panel Loops, or when edges look razor sharp, a Make Boolean Mesh fails or leaves holes, planes look wobbly, Mirror And Weld erased an edit, or ZRemesher ruins a hard-surface part.
Write correct Sarvam Saaras STT code for 23 Indic languages—REST modes, Batch API with diarization, and WebSocket streaming gotchas. Use this skill when building transcription or voice apps in Python or JS/TS. For live transcription in chat via MCP, use sarvam-mcp instead
Build or review a Stripe integration with current Stripe guidance: Checkout Sessions, Payment Element, Setup Intents, subscriptions and the Customer Portal, webhook-driven fulfillment, idempotency, restricted API keys, and Stripe Tax. Use for: stripe, payments, checkout, subscription, billing, webhook, refund, payment intent
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.