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. 79 354 files from 1 739 authors, of which 61 713 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.
| Researches meeting attendees and their companies before any meeting using real-time web data. Surfaces roles, recent activity, company context, and talking points — then maps cross-attendee relationships. Use this skill when the user asks to prepare for a meeting, research someone for my meeting", "who am I meeting with", "research this person", "meeting prep", "brief me on [person]", "I have a meeting with [person/company]", "get me ready for my call", "what should I know about [person]", "background on [person] before our meeting", "attendee research". Requires the Nimble CLI (nimble search, nimble extract) for live web data. Do NOT use for multi-company competitor monitoring (use competitor-intel) or single-company deep dives without attendees (use company-deep-dive).
| Get web data now — fast, incremental, immediately responsive to what the user needs. The only way Claude can access live websites. Must be pre-installed and authenticated. Run `nimble --version` to verify (>= 1.2.0).
| keyword research, rank tracking, site audits, content gap analysis, competitor keyword reverse-engineering, AI visibility across five platforms (ChatGPT, Perplexity, Google AI, Gemini, Grok), and GitHub repo SEO. Crawls real sites and SERPs via Nimble CLI — no fabricated metrics. "competitor keywords", "AI visibility", "GitHub SEO", "SERP analysis", "keyword research", "technical SEO", "keyword difficulty", "topic clusters", "ranking delta", "on-page SEO", "AI citation audit". Do NOT use for competitor business signals — use `competitor-intel` instead. Do NOT use for competitor messaging — use `competitor-positioning` instead. Do NOT use for general web scraping — use `nimble-web-expert` instead.
| Reference for Nimble Web Search Agents (Agent API V2). Load when a task needs open-ended research, enrichment, or dataset building — where the source isn't fixed, data is scattered, structure is inconsistent, or a synthesized answer is needed. locking, run-level `skill` override, run controls, live events vs polling, trust and citations, and safe credentials.
| Reference for nimble crawl command. Load when bulk-crawling many pages asynchronously.
| Reference for nimble extract command. Load when fetching URLs or scraping pages. parser schemas, geo targeting, async, parallelization.
| Reference for nimble map command. Load when discovering URLs on a site before bulk extraction. response structure {links[].url/title/description}, map→filter→extract pattern, map vs crawl comparison.
| Reference for nimble tasks and batches commands. Load when polling async task status, tracking batch progress, or fetching results. crawl (per-page tasks), search async, map async.
| Reference for nimble search command. Load when searching the live web. search_depth modes (lite/fast/deep), response structure, credit costs.
Create narrated video overviews from source documents using editable storyboard rows, OpenAI text-to-speech audio, OpenAI image generation, captions, and deterministic ffmpeg assembly. Use when Codex needs to turn PDFs, DOCX files, Markdown, text notes, research, meeting materials, briefs, or folders of documents into a NotebookLM-style static-image video overview with voiceover, or when the user wants to edit/regenerate individual narration or image rows without rebuilding the whole project.
Build lightweight mood boards, visual territories, creative directions, style boards, reference boards, art-direction boards, campaign mood explorations, brand look-and-feel explorations, image batches for visual ideation, more-like-this rounds from attached images, and follow-up mood-board iterations from annotated or attached images. Use this standalone skill instead of Creative Production when the user asks Codex to create, generate, explore, iterate, or review mood-board-style image directions.
Record and live-transcribe a meeting or live conversation from a Codex thread on macOS, saving streaming transcript text to the workspace with OpenAI Realtime transcription, then summarize notes in-thread. Use when the user asks to start, stop, capture, transcribe, ask questions about, or summarize a meeting recording from the current thread.
Launch and operate a local Codex drawing workspace: a lightweight vector board where the user and Codex share editable scene state through a browser UI, workspace-backed JSON, local API mutations, and SVG export. Use when the user wants to draw, edit, inspect, generate, refine, or export simple vector artifacts such as logos, icons, diagrams, badges, layout sketches, SVG assets, or visual scratchpad scenes.
Generate NotebookLM-style podcasts from documents or supplied scripts. Use when Codex needs to turn source material into a reviewable single-host, two-host, or multi-host podcast script; create host-marked dialogue with stable IDs; manage client revisions and pronunciation guidance; generate ElevenLabs v3 Text to Dialogue audio in natural chunks; or merge chunked audio with FFmpeg into a final episode.
Create polished executive or client presentation decks in Codex through a staged workflow: optional research, user direction, narrative outline, subagent critique, wireframe, mood boards, and final full-slide image design. Use when the user asks to make a polished presentation, turn research, notes, or dictated direction into a deck, create an executive/client deck, build a deck with mood boards, or design slides using image generation. Requires the imagegen skill for mood boards and final slide images.
Persistent Obsidian-based memory for coding agents. Use at session start to orient from a knowledge vault, during work to look up architecture/component/pattern notes, and when discoveries are made to write them back. Activate when the user mentions obsidian memory, obsidian vault, obsidian notes, or /obs commands. Provides commands: init, analyze, recap, project, note, todo, lookup, relate.
Author and optionally direct-launch one durable loop-graph run for long-horizon agent work. Use when the user invokes /longgraph (or legacy /octopus), asks to design or start a multi-round run, needs durable state, gated milestones, owner decision boundaries, cross-host continuity, or independent audit. Detect Codex or Claude Code from context instead of asking. For a self-contained task, recommend the host's ordinary task or goal directly. Do not execute or resume existing runtime node files from this authoring skill.
Scaffold a new context-mill skill that walks an agent through an ordered series of narrowly-scoped reference files. Works for any number of steps. Use when the user wants to add a docs-only skill where the agent must read each step in strict order, one at a time, with each step isolated, with clear endpoints, and no prefetching.
Cut version bumps + changelogs for maintained packages in this monorepo. Detects which packages changed since their published npm version (git-vs-npm), reconciles against pending Changesets, runs `changeset version` to bump + cascade dependents, and opens a PR. This repo only produces version-bump commits; the npm publish is then triggered manually from a separate Databricks releases repo. Use when asked to "release", "cut a release", "bump versions", or "check what needs releasing".
> Compares the seller's own Mercado Livre listing against the available competing listings of the same catalog product — the buy-box competition — and says where they win, lose, and what to fix first. Use it when a seller asks why they don't win the buy-box, why they don't sell despite the same product, or how they rank inside a catalog, given a Mercado Livre or JoomPulse link or a listing or catalog product identifier. It scores each parameter — price, free shipping, Full, listing type, seller reputation, official store — as Melhor, Na média, Pior, or Criticamente pior, presents reviews and rating as context, and returns a verdict, a comparison table, prioritized actions, and a spreadsheet. Triggers include "why don't I win the buy-box" and the pt-BR "por que não ganho o buy-box". Mercado Livre (Brasil) only; sales and revenue are JoomPulse estimates, not real transactions. For a product and its competitors, use the single-product analysis skill; for one over time, the change-monitor skill.
> Finds new, already-selling product listings inside one Mercado Livre (Brasil) category on JoomPulse — recent listings (by default under about 30 days on air) that are well rated (rating above 3) and have real traction (more than about 30 estimated sales in the last month), all thresholds overridable — and returns them as a product table with a JoomPulse link per listing. Use it when a seller category", "what's launching and already selling", "recent best-sellers in my category", and the pt-BR equivalents "produtos novos na categoria", "novos anúncios que já vendem", "lançamentos em alta na minha categoria". Mercado Livre (Brasil) only; sales and revenue are JoomPulse estimates, not real transactions, while price, rating, and review count are real history. For weak-rated but well-selling products to beat, use the high-demand low-quality skill; for tracking one product over time, use the product change-monitor skill.
> Finds the fastest-growing deep sub-categories (leaf categories, deeper than the third level) under a chosen Mercado Livre (Brasil) category, ranked by month-over-month revenue growth, using JoomPulse market data. Use it when a seller picks a category and wants the niches inside it that are growing fastest growing in this category", "deep growing categories", and the pt-BR equivalents "subcategorias em crescimento", "nichos que mais crescem nesta categoria", "categorias profundas em alta". Sales and revenue are JoomPulse estimates, not real transactions. For one category's opportunity index, use the category-opportunity-index skill; for comparing category aggregates against a user-supplied previous snapshot, use the category-monitor skill; for new products inside a category, use the new-growing-products-in-category skill.
> Finds Mercado Livre (Brasil) products in one category that have high demand but a low rating — listings that sell well yet score at or below a rating threshold the user chooses, an opening to enter with a better offer. Use it when a seller wants weak-rated but well-selling products to beat. The skill asks for a category and a rating threshold, then returns the demand low rating products", "low quality opportunities", "products I can beat", and the pt- BR "produtos com muita demanda e nota baixa", "oportunidades de baixa qualidade", "produtos mal avaliados que vendem". Sales and revenue are JoomPulse estimates, not real transactions; price, rating, and reviews are real history. For brand-new listings use the new-growing- products-in-category skill; for niches without strong incumbents use the uncontested-niche- finder skill; for one product and its competitors use the ml-product-analysis skill.
> Reports the JoomPulse opportunity index (low, medium, or high) for one Mercado Livre (Brasil) category, alongside that category's current monthly market stats — estimated GMV, estimated sales, number of active sellers, number of active listings, and average ticket — then writes a plain-language read on how attractive the category is to enter. Use it when a seller wants a category worth entering", "show the opportunity index for this category", "how big is this market", and the pt-BR equivalents "qual o índice de oportunidade", "vale a pena entrar nessa categoria", "tamanho de mercado da categoria". Sales and revenue are JoomPulse estimates, not real transactions. For ranking the sellers in a category, use the top-sellers-in-category skill; for trending search terms, use the top-keywords-in-my-category skill; for comparing category aggregates against a user-supplied previous snapshot, use the category-monitor skill.
> Analyzes one Mercado Livre product and its competitors. Use this skill when a user wants to size up a single product or the products that compete with it — by a Mercado Livre link, a JoomPulse link, a photo, or a row of data. It returns a product card for the subject plus a ranked table of comparable products, each with price, estimated monthly sales and revenue, logistics, and does this sell", "find similar products", "find competing products", and the pt-BR equivalents "analisar este produto", "quanto vende esse produto", "produtos parecidos", "produtos concorrentes", "análise por foto", "análise por link". For "what new products should I add to my store" across a whole catalog, use the assortment gap-analysis skill instead — this skill analyzes a single product the seller already has in hand.
> Finds fast-growing international (imported) products ACROSS ALL Mercado Livre (Brasil) categories, using JoomPulse, and returns them as one product table that includes each item's category, with price, estimated weekly sales and revenue, rating, time on air, shipping, listing type, seller medal, and a JoomPulse link per item. Use it when a seller wants rising imported products international products", "imported products that are taking off", "cross-border winners across all categories", and the pt-BR equivalents "produtos internacionais em alta", "produtos importados crescendo rápido", "produtos internacionais que mais crescem em todas as categorias". Sales and revenue are JoomPulse estimates, not real transactions. For the same search limited to one category, use the single-category international skill.
> Monitors one Mercado Livre (Brasil) category's aggregate health via JoomPulse — estimated sales, number of products and catalog products, active sellers, the seller-medal distribution, and the monopolization level. Each run builds today's snapshot table and offers it for download; to see what changed, the user sends a table from a previous period and the skill shows the metric-by-metric difference. The baseline is whatever table the user this category", "track category sales and sellers", and the pt-BR "monitorar esta categoria", "o que mudou na categoria", "comparar a categoria com o período anterior". Sales and revenue are JoomPulse estimates, not real transactions. To track one named product over time use the product-change-monitor skill; for a one-shot market-size or opportunity snapshot use the category-opportunity-index skill.
> Monitors how one or a few Mercado Livre products or listings changed over a period — by default week over week. Use it when a user wants to track how a named product changes over time, given a Mercado Livre link, a JoomPulse link, a Mercado Livre listing identifier, or a catalog product identifier. It returns week ago for price, rating, and review count, plus the current estimated weekly sales and revenue, logistics, listing type, seller medal, and time on air, with price changes", "what changed this week", "did the price drop", and the pt-BR equivalents "monitorar este produto", "acompanhar este anúncio", "o preço caiu", "variação de avaliações". Sales and revenue are JoomPulse estimates, not real transactions; price, rating, and reviews are real history. For point-in-time product and competitor analysis, use the single-product analysis skill.
> Ranks the brands within one Mercado Livre (Brasil) category by estimated weekly revenue and tracks how each brand's position changes between periods, using JoomPulse. Each run builds today's brand-ranking table and offers it for download; to see how positions moved, the user sends a brand table from a previous period and the skill shows each brand's movement (rose, fell, new, or out). The baseline is whatever table the user supplies — no hidden session memory. Use it to see the top brands in a category and which are rising or falling. the pt-BR "principais marcas da categoria", "ranking de marcas", "quais marcas estão crescendo". Mercado Livre (Brasil) only; sales and revenue are JoomPulse estimates, not real transactions. To rank sellers (shops) rather than brands use the top-sellers-in-category skill; for category market size use the category-opportunity-index skill.
> Finds product listings that appear to represent the same real-world product as a reference item. Use this skill when a user wants to find duplicate listings, match a product across listings, identify identical products by title or URL, compare product photos against marketplace listings, or find the same product on Mercado Livre.
> Finds fast-growing international (imported) products inside ONE chosen Mercado Livre (Brasil) category, using JoomPulse, and returns them as a product table with price, estimated weekly sales and revenue, rating, reviews, time on air, shipping, listing type and seller medal, a JoomPulse link per item, and a pointer to JoomPro for sourcing. Use it when a seller wants the trending products in this category", "fast-growing imported products", and the pt-BR equivalents "produtos internacionais em alta nessa categoria", "produtos importados que mais crescem", "achados internacionais para importar". Sales and revenue are JoomPulse estimates, not real transactions. For the same search across all categories at once, use the all-categories international skill.
> Snapshot tracker for one Mercado Livre (Brasil) seller via JoomPulse — estimated monthly revenue and sales, listing count, sales trend, reputation, seller medal, categories covered, and cancellation rate. Each run builds today's snapshot table and offers it for download; to see what changed, the user sends the seller's table from a previous period and the skill shows the metric-by-metric difference (old → new). The baseline is whatever table the user supplies — no hidden session memory. Use it to track one specific seller over time. "compare this seller with last period", and the pt-BR "monitorar este vendedor", "acompanhar esta loja", "o que mudou nesse vendedor". Sales and revenue are JoomPulse estimates, not real transactions. To rank many sellers in a category and track how that ranking moves, use the top-sellers-in-category skill. Mercado Livre (Brasil) only.
> Ranks the top sellers in one Mercado Livre (Brasil) category by estimated average monthly monthly sales and revenue, 365-day completed sales, cancellation rate, sales trend, brands, product counts, international shipping, and listing-type counts, with a JoomPulse link each. category and it shows each seller's movement (rose / fell / new) plus the biggest movers. revenue", and the pt-BR "principais vendedores da categoria", "maiores lojas da categoria", "quem mais vende nessa categoria". Sales and revenue are JoomPulse estimates, not real transactions. To monitor one seller over time use the seller-overview-tracker skill; to rank brands use the top-brand-position-tracker skill.
> Lists the top trending search keywords for one Mercado Livre (Brasil) category — the real Mercado Livre search-trends ranking of what shoppers look for — with each keyword's rank and how many products compete for it. Use it when a seller wants to know what people search in their niche, which terms to put in titles category", "trending search terms", "what do people search for", "best keywords for my listings", and the pt-BR equivalents "palavras-chave mais buscadas", "termos em alta na categoria", "o que as pessoas pesquisam", "melhores palavras-chave para meus anúncios". The keywords and ranks are real search-trend data, not estimates. For category market size and opportunity, use the category-opportunity-index skill; for ranking the sellers in a category, use the top-sellers-in-category skill.
> Finds products with no brand (unbranded / no-name / generic) in one Mercado Livre (Brasil) category via JoomPulse — an opening to enter with your own brand / private label, since buyers there aren't anchored to a brand name. Asks for a category, lists the active no-brand listings ranked by estimated weekly demand, and returns a product table (price, estimated weekly sales and revenue, rating, reviews, time on air, shipping, listing type, seller brand", "private-label opportunities", and the pt-BR "produtos sem marca", "genéricos que vendem", "oportunidade de marca própria". Sales and revenue are JoomPulse estimates, not real transactions. NOT for ranking brands (use top-brand-position-tracker), brand-new listings (use new-growing-products-in-category), high-demand low-rated products (use high- demand-low-quality-finder), or niches without platinum sellers (use uncontested-niche- finder).
> Finds low-competition niche products on Mercado Livre (Brasil) — active listings in deep sub-categories (deeper than the third level) of a chosen category that have no platinum seller at all, surfaced via JoomPulse. Use it when a seller wants uncontested niches, gaps without strong incumbents, or where to enter without fighting a dominant platinum seller. platinum", "nicho pouco disputado", "subcategorias sem platinum". It returns one table of niche products, each with a competition signal (number of sellers) and a JoomPulse link. Sales and revenue are JoomPulse estimates, not real transactions; price, rating, and reviews are real history. For one product and its competitors use the ml-product-analysis skill; for fast-growing deep categories rather than products use the growing-leaf-category-tracker skill.
Core agent-browser usage guide. Read this before running any agent-browser commands. Covers the snapshot-and-ref workflow, navigating pages, interacting with elements (click, fill, type, select), extracting text and data, taking screenshots, managing tabs, handling forms and auth, waiting for content, running multiple browser sessions in parallel, and troubleshooting common failures. Use when the user asks to interact with a website, fill a form, click something, extract data, take a screenshot, log into a site, test a web app, or automate any browser task.
Manages shadcn components and projects — adding, searching, fixing, debugging, styling, and composing UI. Provides project context, component docs, and usage examples. Applies when working with shadcn/ui, component registries, presets, --preset codes, or any project with a components.json file. Also triggers for "shadcn init", "create an app with --preset", or "switch to --preset".
>- The moi workspace — the web UI the user chats from, extended with agent-authored applets (widgets, views) plus theme & config. Read this FIRST in two cases. (a) A message carries a hidden nothing else says so. (b) The user uses moi vocab — workspace, applet, widget, view, scratchpad, dashboard, or a `moi` command — or asks to build, edit, customize, or theme the workspace UI or its layout.
Release moi-computer to npm — verify locally, hand off to the gated GitHub Actions workflow, then verify the published package. Defaults to a `next` preview; pass `stable` for a `latest` release. Use when the user asks to publish, release, ship a version, or cut a dev preview.
SLASH-COMMAND-ONLY. Invoke ONLY when the user explicitly types the literal `/watch-video` slash command. Never auto-trigger on phrases like "watch this video," "summarize this video," "take notes on this video," or on a bare YouTube URL. For ad-hoc YouTube questions, fetch the transcript directly (YouTube page or yt-dlp) and skim — do not run this heavyweight pipeline.
| Meta Media Buyer — publish, manage, and analyze Meta (Facebook/Instagram) ad campaigns via the Marketing API. Includes guided setup for first-time users. USE WHEN user says "meta ads", "facebook ads", "instagram ads", "publish ads", "create campaign", "launch campaign", "ad performance", "ROAS", "which ad is winning", "ad spend", "campaign metrics", "pause campaign", "resume campaign", "upload ad images", "ad set performance", "creative performance", "set up meta ads", "connect meta ads", "meta ads setup", or needs to publish, manage, or analyze paid Meta advertising campaigns.
One-click starter pack and curator for Codex/agent skills. Use when a beginner wants a high-quality default skills setup across common needs, when the user asks to install useful skills without researching them one by one, or when the installed skills library needs auditing, deduplication, conflict resolution, global-vs-backup selection, restore planning, or a written skills inventory.
Repo-local meta-skill for improving Xiaobai Skills over time. Use after significant project changes, user feedback, install failures, promotion experiments, or repeated workflow friction to update AGENTS.md, capture project-specific gotchas, or propose new repo-local skills. Do not use globally or after every small task.
| to restate the user's intent to align before working, then diagnose the root cause and reuse the domain's mature, established practices before improvising. Clear prompts pass through untouched.
> Query `.belief_map.sexp` before non-trivial code changes to find module boundaries, dependencies, blast radius, and the minimal files to read. Use for architecture discovery, impact analysis, and boundary checks.
This skill should be used when writing Ruby gems following Andrew Kane's proven patterns and philosophy. It applies when creating new Ruby gems, refactoring existing gems, designing gem APIs, or when clean, minimal, production-ready Ruby library code is needed. Triggers on requests like "create a gem", "write a Ruby library", "design a gem API", or mentions of Andrew Kane's style.
Use this agent when you need to create or update README files following the Ankane-style template for Ruby gems. This includes writing concise documentation with imperative voice, keeping sentences under 15 words, organizing sections in the standard order (Installation, Quick Start, Usage, etc.), and ensuring proper formatting with single-purpose code fences and minimal prose. Examples: <example>Context: User is creating documentation for a new Ruby gem. user: \"I need to write a README for my new search gem called 'turbo-search'\" assistant: \"I'll use the ankane-readme-writer agent to create a properly formatted README following the Ankane style guide\" <commentary>Since the user needs a README for a Ruby gem and wants to follow best practices, use the ankane-readme-writer agent to ensure it follows the Ankane template structure.</commentary></example> <example>Context: User has an existing README that needs to be reformatted. user: \"Can you update my gem's README to follow the Ankane style?\" assistant: \"Let...
Use this agent when reviewing code to ensure features are agent-native - that any action a user can take, an agent can also take, and anything a user can see, an agent can see. This enforces the principle that agents should have parity with users in capability and context. <example>Context: The user added a new feature to their application.\\nuser: \"I just implemented a new email filtering feature\"\\nassistant: \"I'll use the agent-native-reviewer to verify this feature is accessible to agents\"\\n<commentary>New features need agent-native review to ensure agents can also filter emails, not just humans through UI.</commentary></example><example>Context: The user created a new UI workflow.\\nuser: \"I added a multi-step wizard for creating reports\"\\nassistant: \"Let me check if this workflow is agent-native using the agent-native-reviewer\"\\n<commentary>UI workflows often miss agent accessibility - the reviewer checks for API/tool equivalents.</commentary></example>
Run comprehensive agent-native architecture review with scored principles
Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends Codex's capabilities with specialized knowledge, workflows, or tool integrations.
Use when needing to look up book content, find a book by title/author, download an ebook, or reference material from a published book. Triggers on book lookups, ebook downloads, "find the book", "get the PDF/EPUB of". Downloads produce PDF/EPUB/MOBI files - use ebook-extractor skill to convert to text.
Use this agent when you need to analyze code changes from an architectural perspective, evaluate system design decisions, or ensure that modifications align with established architectural patterns. This includes reviewing pull requests for architectural compliance, assessing the impact of new features on system structure, or validating that changes maintain proper component boundaries and design principles. <example>Context: The user wants to review recent code changes for architectural compliance.\\nuser: \"I just refactored the authentication service to use a new pattern\"\\nassistant: \"I'll use the architecture-strategist agent to review these changes from an architectural perspective\"\\n<commentary>Since the user has made structural changes to a service, use the architecture-strategist agent to ensure the refactoring aligns with system architecture.</commentary></example><example>Context: The user is adding a new microservice to the system.\\nuser: \"I've added a new notification service that integrates with...
>- Migrate docs from bd (beads) to br (beads_rust). Use when updating AGENTS.md, converting bd commands, "bd sync" → "br sync --flush-only", or beads migration.
Render Mermaid diagrams as readable ASCII/Unicode art in the terminal (from .mmd/.mermaid files, stdin, or Markdown ```mermaid fences). Use when installing or using lukilabs/beautiful-mermaid, creating a CLI renderer for Mermaid-to-ASCII output, previewing Mermaid diagrams in terminal, or extracting/rendering Mermaid blocks from Markdown files.
This skill should be used before implementing features, building components, or making changes. It guides exploring user intent, approaches, and design decisions before planning. Triggers on "let's brainstorm", "help me think through", "what should we build", "explore approaches", ambiguous feature requests, or when the user's request has multiple valid interpretations that need clarification.
This skill should be used when working with AssemblyAI’s Speech-to-Text and LLM Gateway APIs, especially for streaming/live transcription, meeting notetakers, and voice agents that need low-latency transcripts and audio analysis.
Use when user asks to search the web, look something up online, find current/recent/latest information, or needs cited answers. Triggers on "search", "look up", "find out about", "what is the current/latest", image searches, news lookups. NOT for searching code/files—only for web/internet searches.
Use this agent when you need to research and gather external best practices, documentation, and examples for any technology, framework, or development practice. This includes finding official documentation, community standards, well-regarded examples from open source projects, and domain-specific conventions. The agent excels at synthesizing information from multiple sources to provide comprehensive guidance on how to implement features or solve problems according to industry standards. <example>Context: User wants to know the best way to structure GitHub issues for their Rails project. user: \"I need to create some GitHub issues for our project. Can you research best practices for writing good issues?\" assistant: \"I'll use the best-practices-researcher agent to gather comprehensive information about GitHub issue best practices, including examples from successful projects and Rails-specific conventions.\" <commentary>Since the user is asking for research on best practices, use the best-practices-researcher a...
Creates and reviews CLAUDE.md configuration files for Claude Code. Applies HumanLayer guidelines including instruction budgets (~50 user-level, ~100 project-level), WHAT/WHY/HOW framework, and progressive disclosure. Identifies anti-patterns like using Claude as a linter for style rules.
Claude Skills — questions
Answers built from the skills we actually parsed.