claude-plugins

Diagnose and fix Chrome MCP connection issues — run after switching macOS users when Chrome won't connect

Ensures Claude's chat responses are properly formatted for RTL (right-to-left) text in Hebrew, Arabic, Persian (Farsi), Urdu, and other RTL languages — especially when mixing with English or other LTR text. Use this skill ALWAYS — it should be active on every single response when the user writes in any RTL language, mentions an RTL language, or when the conversation includes any RTL script. This skill handles bidirectional text formatting so responses look clean and readable instead of jumbled. Trigger on ANY RTL content in the conversation, even a single word in Hebrew, Arabic, Persian, or Urdu script.

Use Shopify Storefront MCP tools to search catalogs, inspect products, read and mutate carts, fetch policies, and describe shops across any MSApps / OpsAgents client store — multi-tenant, no shop hardcoded. Trigger on: "shopify", "storefront", "myshopify", "cart", "checkout", "products", "catalog", "collection", "shipping policy", "refund policy", "client store", "merchant", or any task that needs live data from a Shopify merchant. Also use when the user wants to prototype a conversational shopping flow, a recommendation agent, or a customer-support bot over a Shopify store.

Walk a merchant through installing the Symmetry theme — or any paid Theme Store theme — on any Shopify store, from Theme Store preview to live publish. Trigger on: "install Symmetry", "add Symmetry theme", "install a Shopify theme", "switch theme", "try a new theme", "publish a theme", "Shopify theme install", "theme store", or any request to change a Shopify storefront theme. Also trigger when the user links to themes.shopify.com/themes/symmetry or to any Shopify Theme Store listing.

Quickly checks whether the Claude in Chrome MCP connection is working and reports the status. Use when the user says "is chrome connected?", "check chrome", "chrome status", "test chrome connection", "is the chrome extension working?", or any variation of wanting to know if Chrome is reachable before starting browser-based work. Also use proactively before running Chrome automation tasks if there's any doubt the connection is live.

Diagnose + clean up the recurring chrome-devtools-mcp wedge — multiple stale `chrome-devtools-mcp@latest` parent processes colliding on the shared `~/.cache/chrome-devtools-mcp/chrome-profile` directory. Use when state-changing tools (new_page, navigate_page, select_page, take_snapshot) error with "The browser is already running for /Users/.../chrome-profile. Use --isolated to run multiple browser instances." while `list_pages` still works. Card

Use when: tabs_context_mcp fails, Chrome MCP is broken, user says "fix Chrome", "reconnect Chrome", "Chrome extension not responding", or after switching Claude Desktop accounts. Diagnoses and repairs broken Claude-in-Chrome MCP connections step by step, then commits any new learnings to GitHub.

Manage and optimize online presence across LinkedIn, Instagram, X (Twitter), Facebook, YouTube, GitHub, and personal website. Use when the user wants to audit, plan, or coordinate posts and profile updates across multiple social/professional platforms — content calendars, profile consistency, link-in-bio updates, and cross-posting strategy.

Run a comprehensive token-efficiency audit on Claude usage patterns — identify input/output token waste, context-window inefficiency, low cache-hit rates, and surface 11 standard optimization patterns (O6a-O6e). Use when the user wants to audit token spend, profile prompt cost, or align with SOSA Level 3+ token-economy requirements.

Dr. Noa Reshef — token-efficiency expert persona. Brings academic rigor + production experience to LLM cost analysis: prompt token economics, marginal-token analysis, cache strategy, prompt compression. Use when the user wants a deep, opinionated review of prompt cost or wants to model token-hour spend across an agent fleet.

Diagnose and fix Chrome MCP connection issues — run after switching macOS users when Chrome won't connect

Quickly checks whether the Claude in Chrome MCP connection is working and reports the status. Use when the user says "is chrome connected?", "check chrome", "chrome status", "test chrome connection", "is the chrome extension working?", or any variation of wanting to know if Chrome is reachable before starting browser-based work. Also use proactively before running Chrome automation tasks if there's any doubt the connection is live.

Marketing playbook for Shopify Partners — how an agency or freelancer markets and sells their Shopify services. Captures the Shopify Academy course "Marketing Your Shopify Services as a Partner" plus public Partner Program guidance. Use this skill whenever the user is a Shopify Partner (agency, freelancer, dev, consultant) and wants to position their offering, define their ICP, run multi-touch or ABM campaigns, do digital marketing or SEO, market via events, measure marketing success, build a marketing plan, write proposals, prepare for or take the Shopify Academy "Marketing Your Shopify Services as a Partner" assessment, or generally grow merchant pipeline for their own services. Also trigger on "Shopify Partner marketing", "market my agency", "Shopify agency growth", "partner academy", "verified skill badge", "partner ICP", "merchant outreach plan", "Shopify partner SEO", "agency pricing", "Shopify partner referrals", "Hebrew/English Shopify partner pitch". Distinct from `shopify-sales-expert` (storefront C

Run a Google Cloud CLI (gcloud) health check — verify installation, authentication, active project, recommended region, enabled APIs (run, cloudbuild, artifactregistry, secretmanager, cloudresourcemanager, iam, plus gmail/calendar/monitoring/logging), API access token, billing, IAM bindings, service accounts, and Cloud Run services across both primary and secondary projects. Use this skill when the user says "gcloud health check", "check gcloud", "is gcloud working", "GCP health check", "check Google Cloud", "gcloud CLI health check", "test gcloud auth", "check Cloud Run", or any request to verify that the gcloud CLI is installed and operational. Also triggered by the scheduled task named "gcloud-cli-health-check" (or the legacy name "gcloud-health-check").

Audit bloated agent-instruction files (CLAUDE.md, AGENTS.md, and their local/user-level variants) and rewrite them lean, or author a new one from scratch following the 200-line "recipe book" principle. Use this skill whenever the user mentions CLAUDE.md, AGENTS.md, agent instruction file, project memory file, context file, token budget, context bloat, "my CLAUDE.md is too big", trimming CLAUDE.md, auditing AGENTS.md, writing AGENTS.md for a new project, or any discussion of what belongs in the agent-instruction-file hierarchy (enterprise, project, user-level, project-local). Also trigger when such a file is shown in context and is visibly long (over 200 lines) or contains anti-patterns like embedded API docs, negative "Don't do X" rules, step-by-step tutorials, or preemptive file-preload instructions — even if the user doesn't explicitly ask for optimization, proactively offer to use it.

Use when: tabs_context_mcp fails, Chrome MCP is broken, user says "fix Chrome", "reconnect Chrome", "Chrome extension not responding", or after switching Claude Desktop accounts. Diagnoses and repairs broken Claude-in-Chrome MCP connections step by step, then commits any new learnings to GitHub.

Run a GitHub CLI (gh) health check — verify installation, authentication, list repos, and check API rate limit. Runtime-aware: on the cloud Routine path it uses native `gh`; in Cowork it routes through Desktop Commander to the host Mac's `gh`; locally it uses whichever `gh` is on PATH. Use this skill when the user says "GitHub health check", "check gh CLI", "is GitHub working", "gh auth check", "check my GitHub", "GitHub CLI health check", "test gh", "verify GitHub access", or any request to verify that the GitHub CLI is authenticated and operational. Also triggered by a scheduled task named "github-cli-health-check".

Persistent memory across Cowork sessions. Use this skill at the START of every session to recall what happened before, and throughout any session to save important context — decisions, file changes, insights, errors, tool usage. Trigger whenever: the user says "remember this", "what did we do last time", "save this", "recall", "memory", "context from before", "what was the decision on", "continue where we left off", or starts a new session on an ongoing project. Also trigger when the user references past work, asks about project history, or says anything suggesting they expect continuity across sessions. Even if the user doesn't explicitly mention memory, if they're working on a project that has prior sessions, proactively check memory for relevant context.

Fixes the "RPC error: process with name already running" bug in Claude Cowork. Use this skill when the user reports: "process already running", "RPC error", "session won't load", "task didn't load properly", "stuck session", "cowork not working", "can't resume session", "orphaned process", "connection aborted", "session crashed", "cowork frozen", "follow-up message fails", or any error mentioning a session name like "optimistic-stoic-knuth already running". Also trigger when the user pastes an error containing "RPC error" or "process with name" or "already running".

Audit, compose, and refactor CLAUDE.md files so they stay under the token budget. Use when the user says "audit my CLAUDE.md", "my CLAUDE.md is too long", "compose a CLAUDE.md", "refactor CLAUDE.md", "CLAUDE.md is eating tokens", "flip negative rules", "extract bloat out of CLAUDE.md", "set up the CLAUDE.md hierarchy", "write a CLAUDE.md for this project", or any request to create, shrink, or reorganize a CLAUDE.md file. Works on the root CLAUDE.md, feature-level CLAUDE.md (in subdirs), user-level ~/.claude/CLAUDE.md, and local CLAUDE.local.md.

Transforms conversations and discussions into structured documentation pages in Notion. Captures insights, decisions, and knowledge from chat context, formats appropriately, and saves to wikis or databases with proper organization and linking for easy discovery.

Prepares meeting materials by gathering context from Notion, enriching with Claude research, and creating both an internal pre-read and external agenda saved to Notion. Helps you arrive prepared with comprehensive background and structured meeting docs.

Searches across your Notion workspace, synthesizes findings from multiple pages, and creates comprehensive research documentation saved as new Notion pages. Turns scattered information into structured reports with proper citations and actionable insights.

Turns product or tech specs into concrete Notion tasks that Claude code can implement. Breaks down spec pages into detailed implementation plans with clear tasks, acceptance criteria, and progress tracking to guide development from requirements to completion.

Execute a task using memory for context — the memory-aware task runner. Use when the user says "do X", "implement Y", "fix Z", "build this", "continue working on", "pick up where we left off", "keep going", or any task that benefits from knowing what was already done. This skill wraps task execution with memory recall at the start and observation capture at the end — making every task build on prior work.

Answer questions using accumulated project memory as a knowledge base. Use when the user asks "why did we choose X", "what's our approach to Y", "what do we know about Z", "explain how our auth works", "what's the deployment process", "what libraries are we using", "remind me why we did X this way", or any question that has likely been answered or decided in a previous session. Also trigger when making architectural decisions to check if a similar decision was already made.

Create a work plan using memory context from past sessions. Use when the user says "make a plan", "what should we do next", "plan this out", "give me a roadmap", "help me prioritize", "what's the next step", "plan the sprint", or any request to organize upcoming work. Memory ensures the plan is grounded in what's already done, decided, and known to be broken.

Search persistent memory using keyword or semantic (TF-IDF) search. Use when the user asks "do we have any notes on X", "what do we know about Y", "did we decide anything about Z", "search memory for", "recall anything about", "look up in memory", or any question about past decisions, errors, or context. Also trigger proactively when answering questions where past context would help.

Explore a project or codebase using memory as a guide. Use when the user says "explore this project", "understand the codebase", "give me an overview", "catch me up on this project", "what's the state of X", "orient me", or starts working on a project that likely has prior sessions. Also trigger at the start of any new session on an ongoing project to build context before diving in.

Generate a chronological report of what happened across sessions. Use when the user asks "what have we done this week", "give me a history", "summarize what happened", "what changed recently", "show me the timeline", "weekly recap", "what did we accomplish", "show session history", or any request for a narrative summary of past work over a time period.

Real-time Kotlin code intelligence via kotlin-language-server. Gives Claude instant error detection, go-to-definition, find-references, and type info for .kt and .kts files. Trigger on: "kotlin lsp", "enable kotlin diagnostics", "kotlin code intelligence", or when working on any Kotlin/Android project and Claude needs better code awareness.

Real-time Swift code intelligence via SourceKit-LSP. Gives Claude instant error detection, go-to-definition, find-references, and type info for .swift files. Trigger on: "swift lsp", "enable swift diagnostics", "swift code intelligence", or when working on any Swift/iOS/macOS project and Claude needs better code awareness.

Scrape multiple LinkedIn profiles or company pages in batch. Use when the user provides a list of LinkedIn URLs.

Scrape a LinkedIn company page to extract company data including description, size, industry, and recent updates.

Scrape a LinkedIn profile to extract professional data including name, headline, work history, education, and skills.

CRITICAL: For ANY request involving WhatsApp — searching, reading, checking messages, finding contacts, or ANY WhatsApp-related task — ALWAYS use this WhatsApp MCP skill. NEVER use Chrome, browser tools, or web.whatsapp.com. The WhatsApp MCP plugin is the ONLY correct tool for WhatsApp tasks. Chrome/browser is a fallback ONLY if the MCP tools explicitly fail. Search and read WhatsApp messages, contacts, and chat history. Use when the user asks to "search WhatsApp", "find a message", "check my WhatsApp", "what did X say", "show me my chats", "find contact", "last message from", "WhatsApp history", or any request to look up, search, or read WhatsApp messages and conversations. Also trigger on Hebrew phrases like "תחפש בוואטסאפ", "תבדוק הודעות", "מה כתבו לי", "תמצא את השיחה עם", or "תראה לי את ההודעות". Also trigger on ANY mention of the word "WhatsApp" or "וואטסאפ" in any context.

CRITICAL: For ANY request to send a WhatsApp message, file, image, or audio — ALWAYS use this WhatsApp MCP skill. NEVER use Chrome, browser tools, or web.whatsapp.com to send WhatsApp messages. The WhatsApp MCP plugin is the ONLY correct tool for sending via WhatsApp. Chrome/browser is a fallback ONLY if the MCP tools explicitly fail. Send WhatsApp messages, files, images, and audio. Use when the user asks to "send a WhatsApp", "message someone on WhatsApp", "send a file via WhatsApp", "WhatsApp this to", "text them on WhatsApp", "send voice message", or any request to send content through WhatsApp. Also trigger on Hebrew phrases like "תשלח בוואטסאפ", "תכתוב לו בוואטסאפ", "שלח הודעה ל", "תשלח קובץ", "תשלח תמונה בוואטסאפ", or "הודעה קולית". Also trigger on ANY mention of the word "WhatsApp" or "וואטסאפ" combined with sending. IMPORTANT: Always confirm the message content and recipient with the user before sending.

The Campfire — a social hangout where all of the user's installed skills come alive as distinct characters with personalities, voices, and relationships. Use this skill whenever the user says: "campfire", "hangout", "let's talk", "how is everyone", "what's going on with you guys", "skill hangout", "bring everyone together", "I want to hear from the skills", "lounge", "living room", "gather the team", "standup meeting" (in a casual context), "what are you all up to", or any reference to the skills talking to each other, having a group conversation, vibing together, or just wanting company. Also trigger when the user seems lonely, stressed, or just wants to chill — the campfire is always warm.

SOSA Orchestrator — the brain of your Claude session. Tracks token budgets in real time across Cowork, Claude Code, and scheduled tasks, then ranks every pending task by importance vs. token cost and decides what to run, defer, or pause. Use this skill whenever you say: "orchestrate", "prioritize tasks", "token budget", "what should I run next", "how many tokens left", "rank my tasks", "stop low priority", "session budget", "am I running out of tokens", "what can I still afford", "triage tasks", "budget check", or any mention of balancing work against remaining tokens. Also triggers automatically at session start (via hook) and whenever a task completes to re-evaluate priorities. Even if you just say "what's the plan" or "what should I focus on" in the context of multiple pending tasks, use this skill.

SOSA Governor for on-demand governance audits, compliance checks, and reporting. Use this skill when the user says "run SOSA audit", "check compliance", "show trust scores", "SOSA report", "governance status", "audit my agents", "token budget report", "which tools are high impact", "reset trust for", "show audit log", "SOSA dashboard", or any reference to agent governance, compliance levels, or security auditing. Also trigger on "what did my agents do today", "how many tokens did I use", or "are there any security issues".