sharkitect-solutions
155 verified skills
meeting-insights-analyzer
Use when analyzing meeting transcripts, recordings, or notes to diagnose dysfunction, power dynamics, facilitation failures, or whether the meeting should exist at all. Use when auditing an organization's meeting culture across multiple meetings. NEVER for scheduling, calendar management, or meeting invitation logistics. NEVER for transcription itself -- this analyzes already-transcribed content.
testing
invoice-organizer
Use when organizing, categorizing, or processing invoices and financial documents, extracting data from PDFs or images of invoices, building invoice filing systems, or preparing financial documents for accounting review. Do NOT use for accounting journal entries, tax filing, financial forecasting, or payment processing.
development
scroll-experience
Use when: user requests scroll-driven animations, parallax storytelling, scroll-triggered reveals, pinned/sticky scroll sections, horizontal scroll layouts, scroll progress indicators, cinematic web experiences, or scroll-based interactive narratives. NEVER for: static page layouts without scroll interaction (use frontend-design), general UI review or component audits (use ui-ux-pro-max), applying preset theme styles (use theme-factory), 3D scene construction (use 3d-web-experience).
development
n8n-expression-syntax
Use when writing n8n expressions in node fields, using {{}} syntax with $json/$node/$now variables, or troubleshooting expressions showing as literal text. NEVER for Code node JavaScript (use n8n-code-javascript) or Python (use n8n-code-python).
development
figma-implement-design
Use when translating Figma designs into production code with pixel-perfect visual parity, implementing specific Figma frames or components from a URL, or validating implemented UI against Figma screenshots. Also use when the user provides a Figma URL with a node ID and asks to build or implement it. NEVER use for designing in Figma, modifying Figma files, general UI development without a Figma source, or Figma MCP setup troubleshooting (use figma skill).
tools
figma
Use when implementing UI from Figma designs, translating Figma nodes to production code, fetching design context or screenshots via Figma MCP, or troubleshooting Figma MCP integration. NEVER use for designing in Figma, non-Figma design tools, or creating designs from scratch without a Figma source.
tools
client-reporting
Use when creating client-facing reports, progress updates, milestone documentation, retainer reports, ROI summaries, or project status dashboards. Use when building QBR decks, writing executive summaries, or designing client-facing KPI dashboards for service businesses. Do NOT use for internal team reports, analytics dashboards for internal use, marketing performance reports (use marketing-demand-acquisition), or financial statements (use smb-cfo).
tools
sales-enablement
When the user wants to create sales collateral, pitch decks, one-pagers, objection handling docs, or demo scripts. Also use when the user mentions 'sales deck,' 'pitch deck,' 'one-pager,' 'leave-behind,' 'objection handling,' 'deal-specific ROI analysis,' 'demo script,' 'talk track,' 'sales playbook,' 'proposal template,' 'buyer persona card,' 'help my sales team,' 'sales materials,' or 'what should I give my sales reps.' Use this for any document or asset that helps a sales team close deals. For competitor comparison pages and battle cards, see competitor-alternatives. For marketing website copy, see copywriting. For cold outreach emails, see cold-email.
development
agent-development
Use when creating Claude Code plugin agents (.md files in agents/ directory), writing agent descriptions for reliable triggering, designing agent system prompts, choosing agent tool sets, selecting agent models (haiku/sonnet/opus), or debugging why an agent doesn't trigger or triggers incorrectly. Use when the question is about Claude Code PLUGIN agents specifically. NEVER for general AI agent architecture (use ai-agents-architect). NEVER for agent memory design (use agent-memory-systems). NEVER for agent evaluation (use agent-evaluation).
tools
agent-evaluation
Use when testing or benchmarking LLM agents before deployment. Use when designing eval suites for agent correctness, reliability, or safety. Use when setting up LLM-as-judge scoring pipelines. Use when debugging why an agent 'feels worse' but metrics don't show it. Use when choosing between human evaluation, automated rubrics, or A/B testing. NEVER for evaluating Claude Code skills specifically — use skill-judge for that.
development
analytics-tracking
Use when implementing analytics tracking, building measurement plans, configuring GA4 or GTM, designing event taxonomies, setting up conversion tracking, or auditing existing tracking implementations. Also use when choosing between analytics platforms, debugging tracking issues, designing UTM strategies, or implementing consent-compliant measurement. NEVER use for A/B test design and statistical analysis (ab-test-setup), CRO experiment recommendations (page-cro, form-cro), SEO performance analysis (seo-optimizer), or marketing attribution modeling (marketing-demand-acquisition).
development
hq-brand-review
Use when reviewing content for Sharkitect Digital brand voice compliance, detecting brand drift across communications, checking for banned terms or tone violations, or conducting a structured brand assessment using the 6-step protocol. Covers all content types: emails, proposals, landing pages, social posts, presentations, and AI-generated outputs. NEVER use for general copywriting feedback without brand focus (use copywriting skill), tone calibration for difficult conversations (use communication-excellence-coach agent), or visual design review (use ui-ux-designer agent).
testing
paid-ads
When the user wants help with paid advertising campaigns on Google Ads, Meta (Facebook/Instagram), LinkedIn, Twitter/X, or other ad platforms. Also use when the user mentions 'PPC,' 'paid media,' 'ad copy,' 'ad creative,' 'ROAS,' 'CPA,' 'ad campaign,' 'retargeting,' or 'audience targeting.' This skill covers campaign strategy, ad creation, audience targeting, and optimization.
development
content-research-writer
Use when writing research-backed articles, blog posts, newsletters, or long-form content requiring source discovery, citation management, and iterative section-by-section refinement. Also use when the user mentions content research, article writing with citations, writing partner, or collaborative drafting. NEVER use for quick social media posts (use social-content), email copy (use email-composer), copy editing of existing text (use copy-editing), or copywriting from scratch without research needs (use copywriting).
testing
ab-test-setup
When the user wants to plan, design, or implement an A/B test or experiment. Also use when the user mentions "A/B test," "split test," "experiment," "test this change," "variant copy," "multivariate test," or "hypothesis." For tracking implementation, see analytics-tracking. For CRO strategy and test ideas, see page-cro. For statistical methods beyond experimentation, see statistical-analysis.
testing
accessibility-auditor
WCAG compliance audits, ARIA implementation, screen reader testing, and accessibility remediation. TRIGGER: "accessibility", "a11y", "WCAG", "ARIA", "screen reader", "keyboard navigation", "ADA compliance", "Section 508", "color contrast", "alt text", "focus management", "assistive technology", "inclusive design", "accessibility audit" EXCLUDE: General UI/UX design (use ui-ux-pro-max), CSS styling without accessibility context, SEO optimization
development
agent-judge
Use when evaluating, auditing, scoring, or reviewing Claude Code native agents (.md files in ~/.claude/agents/) for quality. Use when comparing agents, running agent quality audits, benchmarking agent effectiveness, or deciding which agents need optimization. Use when a user says "score this agent", "evaluate this agent", "audit my agents", "how good is this agent", "rate this agent". Do NOT use for: skill-judge (evaluating SKILL.md files), ultimate-agent-creator (creating new agents), agent-development (structural guidance for plugin agents), agent-evaluation (runtime performance testing).
tools
agent-memory-systems
Use when designing memory architecture for AI agents or chatbots. Use when choosing between conversation buffer, summary, entity, knowledge graph, or vector store memory types. Use when implementing RAG chunking strategies or retrieval pipelines. Use when debugging agent memory failures — forgetting context, inconsistent answers, or retrieving wrong information. Use when planning memory lifecycle (TTL, consolidation, contradiction handling). NEVER for Claude Code's own MEMORY.md file management — that's a separate system.
development
autogpt-agents
Use when the user wants to build autonomous AI agents using AutoGPT Platform, design visual workflow agents, or evaluate AutoGPT against other agent frameworks. Covers platform vs classic architecture decisions, block design, execution model tradeoffs, and production deployment. Do NOT use for: general agent architecture theory (use ai-agents-architect), LangChain agent patterns (use agents-llamaindex or langchain-agent), CrewAI multi-agent orchestration (use agents-crewai), or agent memory system design (use agent-memory-systems).
development
ai-agents-architect
Use when deciding WHETHER to build an AI agent (vs pipeline/chain), choosing an agent architecture pattern (ReAct, Plan-Execute, routing, multi-agent), designing tool schemas for agents, or debugging agent failures (loops, hallucinated tool calls, degraded tool selection). Use when the question is about agent DESIGN, not implementation. NEVER for implementing specific agent frameworks (use agent-development, agents-crewai). NEVER for agent memory design (use agent-memory-systems). NEVER for agent evaluation (use agent-evaluation).
tools
api-patterns
Use when designing REST APIs, choosing between API styles (REST/GraphQL/gRPC/tRPC), implementing authentication patterns, designing rate limiting, versioning APIs, or reviewing API architecture. Do NOT use for Stripe-specific API integration (use stripe-best-practices), MCP protocol design (use mcp-integration), or frontend data fetching patterns unrelated to API design.
tools
app-builder
Use when building a new full-stack application from scratch, scaffolding a project, selecting a tech stack, or coordinating multi-agent application development. NEVER use for single-file scripts, adding features to an existing app (use feature-building.md directly), or non-application artifacts like documents or reports.
development
context7-mcp
Use when answering questions about external libraries, frameworks, SDKs, or package APIs where training data may be outdated or insufficient. Use when a user asks "How do I...", "What's the API for...", "Show me an example of..." involving a specific package. Use when generating code that depends on library-specific syntax, configuration, or API signatures. Use when the user mentions a version number ("React 19", "Next.js 15", "Prisma 6"). Use when you are unsure if your training data reflects the current API surface. Do NOT use for: language builtins or standard library (Python stdlib, JS built-ins -- Claude knows these), conceptual questions not tied to a specific package API, internal or proprietary libraries not in public registries, debugging runtime errors (use systematic-debugging skill).
tools
canvas-design
Use when creating visual art, poster design, programmatic canvas art, PDF/PNG visual compositions, design philosophy-driven artwork, or abstract visual design. NEVER for static page layouts or component design (use frontend-design). NEVER for applying preset theme styles or design tokens (use theme-factory). NEVER for presentation slides (use pptx). NEVER for implementing Figma designs in code (use figma-implement-design).
development
ceo-advisor
Use when the user is a CEO, founder, or executive facing a decision that feels high-stakes, irreversible, or lonely -- hiring/firing a leader, raising capital, managing a board conflict, navigating a crisis, or deciding whether to pivot. NEVER for generic business strategy Claude already knows (Porter's, SWOT, OKRs) or for operational/IC-level work (marketing plans, code reviews, sprint planning).
development
clean-code
Use when writing, reviewing, or refactoring code in any language. Applies to new features, bug fixes, refactoring tasks, and code review. NEVER for documentation-only tasks, data processing one-offs with no maintainability requirement, or throwaway scripts where readability does not matter.
development
competitive-ads-extractor
Use when extracting, analyzing, or comparing competitors' ads from ad libraries (Facebook Ad Library, LinkedIn Ad Library, Google Ads Transparency Center). Also use when the user wants to understand competitor messaging, creative patterns, or ad strategy. NEVER use for creating ads (use copywriting or social-content), running ad campaigns, or non-advertising competitive analysis.
development
contract-legal
Use when drafting, reviewing, or structuring business contracts, service agreements, NDAs, SLAs, terms of service, IP assignments, freelancer agreements, or legal clause selection. Use when building contract templates, negotiating terms, or reviewing counterparty redlines. Do NOT use for regulatory compliance audits, litigation strategy, tax law, employment law disputes, or any matter requiring licensed attorney review.
development
cto-advisor
Use when advising on CTO-level decisions: tech debt prioritization, rewrite vs refactor, build vs buy, engineering team scaling, architecture governance, vendor lock-in assessment, or production reliability strategy. NEVER for individual contributor coding tasks, code review, or project management mechanics.
development
database
Use when designing database schemas, optimizing queries, choosing between database engines, planning data models, implementing migrations, or troubleshooting database performance issues. Do NOT use for Supabase-specific patterns (use supabase-postgres-best-practices), application-level caching strategy, or ORM-specific configuration unrelated to database design.
testing
Data Privacy Compliance
Use when building systems that collect, store, or process personal data and the user needs to determine which regulations apply, choose a lawful basis for processing, structure data processing agreements, handle cross-border transfers, or respond to data breaches. Use when privacy architecture decisions have enforcement consequences. NEVER for general security hardening without a privacy component. NEVER for cookie banner HTML/CSS implementation (use frontend skills). NEVER for HIPAA clinical workflow design (use healthcare-specific skills).
tools
deslop
Use when cleaning up AI-generated code, removing unnecessary comments, defensive checks, redundant type assertions, or style inconsistencies introduced by AI coding assistants. Also use when the user says deslop, clean up AI code, or remove slop. NEVER use for refactoring human-written code, changing application logic, or code review (use clean-code).
development
dispatching-parallel-agents
Use when facing 2+ independent tasks that can run concurrently without shared state. Use when multiple test files fail with different root causes. Use when independent subsystems need investigation simultaneously. Use when parallel code generation, research, or analysis would save time. NEVER when tasks share state, edit the same files, or have sequential dependencies.
development
Email Composer
Use when composing a NEW email from scratch (any type: business, technical, customer, personal-professional). NEVER for editing/polishing existing drafts (email-draft-polish), multi-email sequences or drip campaigns (email-sequence), email infrastructure/deliverability/SPF/DKIM (email-systems), cold sales outreach (cold-email), internal company announcements/memos (internal-comms), general text editing (copy-editing), or non-email professional writing (professional-communication).
devops
error-resolver
Use when encountering any error message, stack trace, exception, or unexpected behavior during development. Use when debugging build failures, runtime crashes, test failures, dependency conflicts, configuration errors, or permission errors. NEVER for performance optimization without an error present, feature requests, code review tasks with no error, or general refactoring work.
development
executing-marketing-campaigns
Use when planning, structuring, or optimizing a marketing campaign -- including channel selection, budget allocation, measurement setup, or diagnosing why a running campaign is underperforming. Also activate when the user asks about go-to-market execution, campaign calendars, or multi-channel coordination. NEVER for brand identity/voice work (use brand skills), pure copywriting without campaign context (use copywriting skills), or ad creative review.
development
executing-plans
Use when you have a written implementation plan to execute in a separate session with review checkpoints
testing
executing-plans
Use when you have a written implementation plan file to execute in a session with human review checkpoints between batches. NEVER use when no written plan exists, when the task is exploratory or open-ended, or when no clear stop-and-report structure has been agreed upon.
testing
i18n-localization
Internationalization architecture, translation management, and locale-aware formatting. TRIGGER: "i18n", "internationalization", "localization", "l10n", "translations", "locale", "RTL support", "right-to-left", "pluralization", "ICU message format", "react-i18next", "next-intl", "vue-i18n", "gettext", "translation keys", "language switching" EXCLUDE: Content writing/copywriting (use copywriting), general React/Vue/Angular questions without i18n context
development
file-organizer
Use when organizing files and directories, cleaning up cluttered folders, establishing filing conventions, detecting and handling duplicate files, or designing folder taxonomy for projects. Do NOT use for invoice/financial document organization (use invoice-organizer), source code project structure (use senior-architect), or database file management.
development
skills/find-bugs
--- name: find-bugs description: Use when asked to review code changes for bugs, find security vulnerabilities, audit code quality on a branch, or perform a security review of local changes. Also triggered by: "review my changes," "check for bugs," "security audit," "code review this branch." NEVER for style-only reviews, formatting checks, or reviewing code that has not been changed in the current branch. version: "2.0" optimized: true optimized_date: "2026-03-10" --- # Find Bugs Review chang
development
free-tool-strategy
When the user wants to plan, evaluate, or build a free tool for marketing purposes -- lead generation, SEO value, or brand awareness. Also use when the user mentions "engineering as marketing," "free tool," "marketing tool," "calculator," "generator," "lead gen tool," or "interactive tool." For viral quiz/generator mechanics, see viral-generator-builder. For SEO page scaling, see programmatic-seo. For demand generation channels, see marketing-demand-acquisition.
tools
game-changing-features
Use when the user wants to identify 10x product opportunities, asks "what should we build next", says "game-changing", "10x", "high-leverage features", or wants to discover transformative product moves rather than incremental improvements. NEVER for RICE scoring or PRD writing (product-manager-toolkit). NEVER for brand positioning or go-to-market strategy (product-strategist). NEVER for marketing campaign ideation (marketing-ideas). NEVER for executive-level business strategy (ceo-advisor).
tools
git-commit-helper
Use when generating commit messages from git diffs, reviewing staged changes for commit readiness, or helping structure multi-file commits into atomic units. Also use when the user asks for help writing commit messages, splitting changes into separate commits, or determining commit scope and type. NEVER use for git workflow strategy (branching, merging, rebasing), git troubleshooting or error resolution, or pull request descriptions and reviews.
tools
hq-orchestrator
Use when a task touches multiple business domains (revenue + brand, tech + operations, legal + finance), requires coordinating 2+ specialist agents, needs cross-department synthesis for executive decision-making, or when the user asks to route a request to the right team/department. NEVER use for single-domain tasks where one specialist agent suffices (use that agent directly), generic task parallelism without business routing logic (use dispatching-parallel-agents skill), or general multi-agent coordination without Sharkitect business context (use multi-agent-coordinator agent).
testing
hq-reverse-engineering
Use when the user wants to reverse engineer a competitor's product or approach, analyze a YouTube video or tutorial to extract architectural insights, deconstruct an existing system to understand how it works, or gather structured competitive intelligence from vague or unstructured sources with confidence-level scoring. NEVER use for standard competitive market research with public data (use competitive-intelligence-analyst agent), general web research without reverse engineering intent (use search-specialist agent), or analyzing your own codebase (use code-reviewer or architect-reviewer agents).
development
hq-tech-strategy
Use when making technology architecture decisions for Sharkitect operations, evaluating build-vs-buy for platforms/tools, assessing security tiering for new integrations, applying the MVA (Minimal Viable Architecture) principle, or when the user asks about tech stack decisions, platform selection, or technology risk assessment. NEVER use for generic software architecture without Sharkitect business context (use senior-architect skill), generic CTO advisory without specific tech decisions (use cto-advisor skill), or AI/LLM system design (use ai-systems-architect agent).
tools
internal-comms
Use when asked to write any internal communication: status reports, leadership updates, 3P updates, company newsletters, FAQ responses, incident reports, project updates, or any message intended for an internal audience. NEVER for external communications (press releases, customer emails, marketing copy, proposals, or public-facing content).
testing
marketing-demand-acquisition
Use when planning demand generation campaigns, optimizing paid media spend across channels, building acquisition strategies, or calculating customer acquisition costs. Also use when the user mentions demand gen, paid ads, LinkedIn ads, Google ads, Meta ads, CAC, acquisition strategy, lead generation, pipeline generation, or MQL/SQL optimization. NEVER use for brand marketing or awareness campaigns without acquisition goals (use marketing-strategy-pmm), content creation or copywriting (use content-creator or copywriting), or email sequence design (use email-sequence).
development
marketing-ideas
Use when user needs marketing ideas, growth tactics, or asks how to promote/market/grow their product. Use when user says 'marketing ideas,' 'growth ideas,' 'how to market,' 'ways to promote,' 'marketing tactics,' or 'what channels should I use.' Also use when user is stuck on which marketing channels to prioritize or needs to evaluate marketing options. NEVER for brand strategy (use marketing-strategy-pmm). NEVER for writing marketing copy (use copywriting or content-creator). NEVER for executing a specific campaign (use executing-marketing-campaigns).
testing
marketing-psychology
Use when applying psychological principles or mental models to marketing copy, pricing, landing pages, or conversion optimization. Also use when analyzing why a marketing approach is or isn't working through a behavioral science lens. NEVER use for clinical psychology, academic research, or non-marketing behavioral analysis.
tools
marketing-strategy-pmm
Use when user needs product positioning (April Dunford method), GTM strategy, competitive battlecards, ICP definition, launch planning, international market entry, or sales enablement content. Use when user mentions positioning, GTM, go-to-market, competitive analysis, battlecard, win/loss, market entry, or product launch. NEVER for demand generation execution, paid media management, content calendar creation, or marketing automation workflows -- those belong to demand-gen skills.
tools
customer-success
Use when managing post-sale customer relationships, building health scores, predicting churn, preparing QBRs, designing renewal playbooks, identifying expansion revenue, creating customer segmentation models, building NPS/CSAT programs, or designing escalation frameworks. Use when analyzing customer lifecycle stages from onboarding through advocacy. Do NOT use for lead generation, prospecting, or pre-sale activities (use lead-research-assistant). Do NOT use for deal closing, proposal writing, or sales pipeline management (use sales-enablement). Do NOT use for product feature development or roadmap planning.
development
market-research-reports
Use when producing market research analysis, market sizing, competitive landscape reports, or investment-grade market assessments. NEVER for product-level competitive feature comparisons, financial modeling/DCF, or general business writing that does not require market data analysis.
testing
copy-editing
Use when editing, reviewing, or improving existing marketing or conversion copy. Also use when the user mentions copy feedback, proofreading, polish, or copy sweep. NEVER use for writing new copy from scratch (use copywriting skill), structural page optimization (use page-cro), or non-marketing content editing.
business
content-creator
Use when creating written content: blog posts, articles, case studies, whitepapers, social media posts, email copy, landing pages, or newsletters. Use when user asks to write, draft, create content, needs copywriting, or mentions content strategy. NEVER for code documentation, technical specs, or API docs -- those are developer-writing tasks, not content creation.
development
competitor-alternatives
Use when creating competitor comparison pages, alternative-to landing pages, vs-competitor content, competitive battle cards, or win/loss analysis documentation. Use when a user says 'alternative page,' 'vs page,' 'comparison page,' 'battle card,' '[Product] vs [Product],' or 'competitor analysis content.' Do NOT use for general content marketing, SEO strategy unrelated to competitive pages, product roadmap decisions, or pricing strategy.
testing
mcp-integration
Use when: add MCP server, configure MCP, .mcp.json, Model Context Protocol, connect external service via MCP, mcp__ tool prefix, MCP debugging, MCP server not connecting. NEVER for: Claude Code hooks or event triggers (use hook-development), general REST API development without MCP, webhook endpoint creation, plugin.json structure beyond mcpServers field (use plugin-dev:skill-development).
tools
cold-email
Use when writing cold outreach emails, prospecting emails, cold email campaigns, sales development emails, SDR emails, or follow-up sequences for cold prospects. Also use when user asks to personalize cold emails, write subject lines for cold outreach, or build multi-touch cold email sequences. NEVER for lifecycle/nurture emails (use email-sequence), landing page copy (use copywriting), or LinkedIn/social outreach (use social-content).
development
templates
Project scaffolding templates for new applications. Use when creating new projects from scratch. Contains 12 templates for various tech stacks.
data-ai
n8n-code-javascript
Use when writing JavaScript in n8n Code nodes, choosing between Code node and built-in nodes, debugging Code node errors (return format, undefined data, silent failures), or implementing pagination/retry/dedup in code. NEVER for expression fields (use n8n-expression-syntax) or Python Code nodes (use n8n-code-python).
development
n8n-code-python
Use when writing Python in n8n Code nodes, using _input/_json/_node syntax, or choosing between Beta and Native Python modes. NEVER for JavaScript Code nodes (use n8n-code-javascript) or expression fields (use n8n-expression-syntax).
development
document-lifecycle
Use when (1) a hook signals potential drift between current work and tracked documents, (2) session-start alerts show documents due/overdue/critical for review, (3) user requests a document freshness check, or (4) completing a task that changed business direction, pricing, processes, or strategy. Two modes: drift resolution (compare doc against current work, present contradictions) and scheduled review (read doc, diff against recent activity, ask targeted confirmation questions). Manages escalation ladder from gentle reminder through hard stop. Do NOT use for: content creation enforcement (use hq-content-enforcer), post-task resource gap detection (use resource-auditor), brain memory management (handled by supabase-sync).
testing
n8n-workflow-patterns
Use when designing n8n workflow architecture, choosing between webhook/API/database/AI/scheduled patterns, configuring AI agent sub-nodes, or troubleshooting responseMode and execution order issues. NEVER for individual node configuration (use n8n-node-configuration) or MCP tool usage (use n8n-mcp-tools-expert).
tools
nextjs-best-practices
Use when building Next.js applications with App Router, designing server/client component architecture, implementing data fetching strategies, configuring caching and revalidation, or optimizing Next.js performance. Do NOT use for Pages Router legacy patterns, general React questions unrelated to Next.js, or static site generators other than Next.js.
tools
notion-template-business
Build and sell Notion templates as a sustainable digital product business. Use when: (1) user wants to create Notion templates for sale, (2) user asks about template pricing or packaging, (3) user wants marketplace strategy for Notion templates, (4) user needs help scaling template revenue. Do NOT use for: general Notion workspace setup (use notion skills), generic digital product advice without Notion focus (use micro-saas-launcher), copywriting for sales pages (use copywriting), pricing theory without Notion context (use pricing-strategy).
development
onboarding-cro
Use when optimizing post-signup user activation, first-run experience, time-to-value, onboarding checklist design, empty state optimization, or activation rate improvement. Also for stalled user re-engagement, onboarding email triggers, or feature discovery sequencing. NEVER for signup/registration form optimization (use signup-flow-cro), ongoing lifecycle email sequences beyond onboarding (use email-sequence), paywall or upgrade flow optimization (use paywall-upgrade-cro), general page conversion optimization (use page-cro).
development
outreach-specialist
Use when planning cold outreach campaigns, optimizing email deliverability, building prospect targeting strategies, managing sending infrastructure, designing follow-up sequences, or analyzing outreach performance. Use when the user mentions cold email, outreach, prospecting, lead generation campaigns, email warm-up, deliverability, or reply rates. NEVER for warm introductions or referral-based outreach. NEVER for email marketing to opted-in lists. NEVER for content marketing or social selling. NEVER for transactional email systems.
development
gemini-api-dev
Use this skill when building applications with Gemini models, Gemini API, working with multimodal content (text, images, audio, video), implementing function calling, using structured outputs, or needing current model specifications. Covers SDK usage (google-genai for Python, @google/genai for JavaScript/TypeScript, com.google.genai:google-genai for Java, google.golang.org/genai for Go), model selection, and API capabilities.
development
paywall-upgrade-cro
When building or optimizing in-app paywalls, upgrade modals, feature gates, usage limit screens, or trial expiration flows. Also when diagnosing why upgrade conversion is low despite product engagement. For public pricing pages use page-cro instead. For non-upgrade popups use popup-cro instead. For form field optimization use form-cro instead. For onboarding-to-aha flows use onboarding-cro instead. For experiment infrastructure use ab-test-setup instead.
development
pdf-processing-pro
Use when extracting text/tables/forms from PDFs, filling PDF forms, OCR on scanned documents, batch PDF processing, diagnosing PDF extraction failures, choosing between PDF libraries. NEVER for Word documents (use docx), spreadsheets (use xlsx), presentations (use pptx), generating PDFs from scratch (use reportlab/weasyprint directly), image-only processing without PDF context (use standard PIL/OpenCV).
documentation
plugin-judge
Use when evaluating, auditing, scoring, or reviewing Claude Code plugins (installed packages containing hooks, skills, agents, MCPs, and scripts) for quality. Use when comparing plugins, running plugin quality audits, deciding which plugins to keep/remove/absorb, or assessing custom plugin builds. Use when a user says "score this plugin", "evaluate this plugin", "audit my plugins", "should I keep this plugin", "rate this plugin". Do NOT use for: skill-judge (evaluating standalone SKILL.md files), agent-judge (evaluating standalone agent .md files), plugin-dev (building/developing plugins), hook-development (writing hooks).
tools
hq-operations
Use when creating or auditing Sharkitect Digital SOPs, detecting operational drift from documented procedures, managing capacity state assessments, or enforcing the 10-field SOP template standard. Covers process documentation, operational health monitoring, and capacity planning. NEVER use for project planning or timeline management (use project-manager agent), sprint facilitation or agile ceremonies (use scrum-master agent), or general business analysis without operational focus (use business-analyst agent).
testing
popup-cro
Use when optimizing popups, modals, overlays, slide-ins, exit intent popups, announcement banners, lead capture popups, popup timing/trigger strategy, popup frequency rules, or overlay conversion rate. NEVER for optimizing form fields inside a popup (use form-cro). NEVER for page-level conversion optimization beyond the popup (use page-cro). NEVER for A/B test statistical design or sample sizing (use ab-test-setup). NEVER for post-conversion email sequences (use email-sequence).
testing
pricing-strategy
Use when the user needs help choosing a pricing model, setting price points, structuring tiers, planning a price increase, or diagnosing pricing-related churn. Also use for packaging decisions, freemium vs trial trade-offs, and value metric selection. NEVER for one-time product pricing (retail, physical goods). NEVER for financial modeling or revenue forecasting -- use spreadsheet tools for that.
tools
product-manager-toolkit
Use when the user needs feature prioritization (RICE scoring, backlog ranking, capacity planning), customer interview analysis (transcript parsing, insight extraction, cross-interview synthesis), PRD writing or review, or product discovery frameworks (opportunity mapping, hypothesis testing). NEVER use for marketing strategy or go-to-market campaigns (use marketing-strategy-pmm). NEVER use for project execution, sprint planning, or delivery tracking (use executing-plans). NEVER use for standalone user research methodology beyond interview transcript analysis.
tools
product-strategist
Use when making product strategy decisions -- prioritization, OKR design, roadmap planning, product-market fit assessment, or build-vs-buy evaluation. Use when diagnosing product health (retention curves, engagement ratios, revenue metrics). Use when translating company vision into team-level execution. Use when a product leader asks 'what should we build next?' or 'how do we know if we have PMF?' NEVER for project management or sprint planning -- that is execution, not strategy. NEVER for marketing positioning or go-to-market -- use marketing skills instead.
development
professional-communication
Use when drafting high-stakes business messages (layoffs, escalations, disagreements with leadership, budget requests, bad news upward). Use when the user's communication problem is political, cross-cultural, or audience-calibration -- not just 'make this sound professional.' NEVER for casual messages, basic email formatting, or grammar fixes.
development
project-lifecycle
Use when starting a new project or feature that requires multi-phase planning, when completing a phase and needing structured cleanup, when a plan changes mid-execution and artifacts need updating, when something fails and the failure needs documenting to prevent repetition, or when ending a session with work in progress. Use when any project has more than 3 tasks or spans multiple sessions. Do NOT use for: single-file edits, trivial changes, or exploratory research without deliverables (use writing-plans for plan creation, executing-plans for plan execution mechanics, systematic-debugging for bug investigation).
development
ultimate-skill-creator
Use when creating a new skill for Claude Code — whether from scratch, from a user request, or by formalizing behavior that should be consistent. Use when a user says "create a skill", "write a skill", "build a skill", "make a skill for", or describes behavior they want Claude to follow reliably across sessions. Use when improving, refactoring, or debugging an existing skill that fails to trigger, gets ignored under pressure, or produces inconsistent results. Use when evaluating whether a skill actually changes Claude's behavior versus baseline. Use when packaging a skill for sharing or marketplace distribution. Use when the task seems "simple" or "quick" — those are the highest-risk cases where Claude skips quality steps.
development
referral-program
Use when designing referral programs, affiliate programs, or word-of-mouth growth loops. Also use when the user mentions referral incentive structure, viral coefficient modeling, affiliate commission design, referral fraud prevention, or ambassador programs. NEVER use for email nurture sequences (email-sequence), psychological persuasion principles (marketing-psychology), pricing model design (pricing-strategy), or launch execution planning (launch-strategy).
testing
security-best-practices
Use when performing security reviews, writing secure-by-default code, generating vulnerability reports, or hardening existing codebases. Also use when the user mentions security audit, secure coding, OWASP, vulnerability assessment, or security best practices. NEVER use for general code review without security focus (use clean-code), penetration testing execution, or compliance certification (use data-privacy-compliance).
development
senior-architect
Use when designing system architecture, evaluating technology trade-offs, creating architecture diagrams, defining service boundaries, planning database schemas, or reviewing architectural decisions. NEVER use for UI/UX design decisions, individual code-level refactoring, or DevOps pipeline configuration.
development
smb-cfo
Use when analyzing P&L statements, building cash flow forecasts, calculating runway or break-even, creating budgets, running monthly close, categorizing expenses, modeling unit economics, or making financial decisions for small-to-medium businesses. Use when revenue is under $50M and the team lacks a full-time CFO. Do NOT use for investment banking models, personal finance, enterprise treasury management, or public company SEC reporting.
development
supabase
Use when doing ANY task involving Supabase products: Database (Postgres + RLS), Auth (sessions, JWT, getUser/getSession/getClaims, MFA, OAuth, magic links), Edge Functions (Deno runtime, secrets, cron, webhooks), Realtime (postgres_changes, broadcast, presence), Storage (buckets, signed URLs, transformations), Vectors (pgvector, HNSW, IVFFlat), Cron (pg_cron), Queues (pgmq); client libraries supabase-js and @supabase/ssr in Next.js / React / SvelteKit / Astro / Remix / React Native; Supabase CLI (supabase migration, db query, db advisors), Supabase MCP server, Postgres extensions (pg_graphql, pg_net, pg_vector, vault). Do NOT use for: deep Postgres query/index/schema optimization (use supabase-postgres-best-practices skill), Realtime WebSocket-level scaling and connection-drop debugging (use supabase-realtime-optimizer agent), generic database schema design without Supabase context (use database skill), generic backend architecture without Supabase (use senior-backend skill).
tools
supabase-postgres-best-practices
Use when designing Supabase database schemas, writing Postgres queries in Supabase projects, configuring Row Level Security policies, optimizing query performance, setting up connection pooling, or troubleshooting Supabase database issues. Do NOT use for general backend architecture, frontend Supabase client usage unrelated to database, Supabase Auth configuration, or Supabase Storage/Realtime features.
tools
testing-strategy
Use when designing test architecture, choosing what to test, building test pyramids, planning E2E test strategy, managing test data, optimizing CI test pipelines, or establishing testing standards for a project. Do NOT use for writing individual unit tests, debugging test failures, or test framework API reference.
development
knowledge-management
Use when designing knowledge bases, creating SOPs, writing runbooks, structuring documentation systems, building internal wikis, or establishing information architecture for teams. Use when someone says 'create a knowledge base', 'write an SOP', 'build a runbook', 'organize our documentation', 'set up an internal wiki', or 'our docs are a mess'. Use when designing tagging taxonomies, content governance models, documentation review cadences, or knowledge transfer plans. Use when migrating between documentation platforms. Use when capturing institutional knowledge during incidents or employee departures. Do NOT use for external-facing product documentation, API documentation, marketing content libraries, or developer documentation portals (use technical-writing or docs-as-code skills instead).
development
launch-strategy
Use when planning a product launch, feature announcement, go-to-market release, or phased rollout. Also use when the user mentions Product Hunt, early access, beta launch, waitlist, launch momentum, or release strategy. NEVER use for ongoing marketing campaigns (use marketing-strategy-pmm), creating launch assets or copy (use copywriting or content-creator), or post-launch analytics and attribution.
development
lead-research-assistant
Use when the user asks to find leads, prospects, target accounts, or potential customers for any product or service. Use when building outreach lists, qualifying prospects, or researching companies for sales or BD. Use when the user mentions ICP, lead scoring, prospect research, or account-based targeting. NEVER for recruiting/hiring research. NEVER for competitor analysis without a lead-gen angle. NEVER for market sizing alone.
development
micro-saas-launcher
Use when user is building a solo/small-team SaaS product targeting a niche market with <$50K MRR goal, needs help with go/no-go decisions, MVP scoping to weeks not months, initial pricing architecture, or launch channel selection for indie/bootstrapped products. NEVER for full application architecture (use app-builder), enterprise or funded launch campaigns (use launch-strategy), advanced pricing optimization or experimentation (use pricing-strategy), market positioning or competitive strategy (use product-strategist), landing page design or CRO (use page-cro).
development
nestjs-expert
Use when building NestJS applications, designing module architecture, configuring dependency injection, implementing guards/interceptors/pipes, or troubleshooting NestJS-specific errors. Do NOT use for general Node.js/Express questions, frontend frameworks, or database queries unrelated to NestJS ORM integration.
development
hr-people-ops
Use when writing job descriptions, designing hiring pipelines, building structured interviews, creating offer letters, planning employee onboarding, setting up performance review cycles, designing compensation frameworks, writing PIPs, planning team restructures, drafting employee handbook sections, or handling offboarding. Use when the context is SMB operational HR (1-200 employees). Do NOT use for enterprise HRIS platform implementation, benefits administration platform configuration, labor law litigation, immigration law, union negotiations, or payroll tax compliance.
development
n8n-validation-expert
Use when interpreting n8n validation errors, choosing validation profiles, fixing operator structure issues, recognizing false positives, or recovering from broken workflow states. NEVER for expression syntax errors (use n8n-expression-syntax) or node configuration (use n8n-node-configuration).
tools
n8n-mcp-tools-expert
Use when working with n8n MCP tools (search_nodes, get_node, validate_node, create/update workflows), choosing nodeType formats, selecting validation profiles, or deploying templates. NEVER for expression syntax (use n8n-expression-syntax) or Code node content (use n8n-code-javascript/python).
tools
pptx
Use when creating new PowerPoint presentations from scratch, editing existing .pptx files, working with slide templates, extracting or replacing slide text, analyzing slide layouts, or generating thumbnail grids. Use when the user mentions .pptx, PowerPoint, slides, or presentations. NEVER for PDF-only workflows, slide conversion to images as the primary goal, or reading presentation text when markitdown alone suffices.
development
theme-factory
Use when applying visual themes (colors, fonts) to slides, documents, reports, or HTML pages, or when creating custom color/font palettes for any artifact. NEVER use for content creation, layout design, or choosing between artifact types.
development
airtable-expert
Use when building, modifying, or planning Airtable bases programmatically -- specifically: creating schemas via REST API, hitting a 422/INVALID_REQUEST error from the Airtable API, writing rollup or formula field expressions, designing interfaces (Dashboard / Record Review / Grid), scoping permissions for client-facing bases (interface-only vs full-editor), or modeling parent-child/many-to-many relationships. Triggers on phrases like 'Airtable base', 'create Airtable table', 'Airtable API', 'rollup formula', 'Airtable interface', 'base schema'. Do NOT use for: general REST API design (use backend-architect), SQL database schema (use database-architect), Notion or other non-Airtable no-code platforms (use the relevant platform skill), or generic spreadsheet design without Airtable-specific features (use xlsx skill).
tools
subagent-driven-development
Use when executing an implementation plan by dispatching subagents for each task, when you have independent tasks to implement from a plan, or when you need fresh-context execution with two-stage review (spec compliance then code quality). Also use when asked to 'implement this plan' or 'execute these tasks'. NEVER for creating the plan itself (use superpowers:writing-plans), for tightly coupled tasks requiring shared state, or for parallel-session execution (use superpowers:executing-plans).
development
telegram-bot-builder
Use when building Telegram bots, designing bot conversation flows, choosing bot libraries, configuring webhooks or polling, implementing Telegram Payments or Stars, handling Telegram API rate limits or file size constraints, or scaling a bot beyond hobby usage. NEVER for voice-based bot experiences (use voice-agents), no-code bot automation via n8n (use n8n-workflow-patterns), full web applications that happen to include a Telegram bot (use app-builder), Discord/Slack/WhatsApp bots (platform-specific skills).
tools
senior-devops
Production infrastructure, CI/CD pipelines, container orchestration, and cloud operations. TRIGGER: "deploy", "infrastructure", "terraform", "kubernetes", "docker compose", "CI/CD", "monitoring", "alerting", "SLO", "load balancer", "autoscaling", "helm", "IaC", "container orchestration", "incident response", "rollback", "blue-green", "canary" EXCLUDE: GitHub Actions YAML (use github-actions-creator), Docker basics for development (use docker-expert), cloud pricing/billing questions
development
prompt-engineering-guidance
Use when building constrained LLM generation with Microsoft Guidance library. Use when outputs MUST match a specific format (JSON, dates, enums, code). Use when choosing between Guidance, Instructor, Outlines, or native JSON mode for structured output. Use when debugging constrained generation failures (slow grammar compilation, quality degradation from overly strict constraints). NEVER for general prompt engineering or prompt writing — this is specifically for the Guidance library and constrained generation patterns.
development
email-systems
Use when building transactional or marketing email infrastructure, configuring SPF/DKIM/DMARC, debugging deliverability issues, setting up email queues with retry logic, or designing email automation workflows. NEVER use for writing email copy (use email-composer or email-draft-polish), designing email templates visually, or non-email notification systems.
tools
email-draft-polish
Use when reviewing a generated cold email draft, polishing email copy before sending, checking an email against brand guidelines, auditing email quality, or validating that an email meets formatting and content rules. Also triggered by: 'review this email,' 'check this draft,' 'polish this email,' 'QA these emails,' 'audit email quality.' NEVER for writing emails from scratch, general copywriting tasks, or non-cold-email correspondence.
development
social-content
Use when creating social media posts, threads, carousels, or platform-specific content for LinkedIn, Twitter/X, Instagram, TikTok, or Facebook. Use when the user mentions social media, content calendar, social scheduling, engagement strategy, viral content, LinkedIn post, Twitter thread, or social repurposing. NEVER for paid social ads or sponsored content (use ads skill). NEVER for email newsletters (use email-sequence). NEVER for blog posts or long-form articles (use content-creator).
testing
ultimate-agent-creator
Use when creating a new Claude Code native agent (.md file in ~/.claude/agents/), when building an agent from scratch or from a user request, when formalizing behavior into a reusable agent, or when the user says "create an agent", "build an agent", "make an agent for", "write an agent". Use when improving, refactoring, or rebuilding an existing agent that fails to trigger, produces generic output, or has over-permissioned tools. Use when optimizing an agent's description for better triggering, rewriting a bullet-list body into expert content, or scoping tools to least-privilege. Do NOT use for: agent-judge (evaluating/scoring existing agents), agent-development (plugin agent structural guidance), skill-judge (scoring SKILL.md files), ultimate-skill-creator (creating skills, not agents).
tools
ultimate-skill-creator
Use when creating a new skill for Claude Code -- from scratch, from a user request, or by formalizing behavior that should be consistent. Use when a user says "create a skill", "write a skill", "build a skill", "make a skill for", or describes behavior they want Claude to follow reliably across sessions. Use when improving, refactoring, or debugging an existing skill that fails to trigger, gets ignored under pressure, or produces inconsistent results. Use when packaging a skill for sharing or marketplace distribution. Use when the task seems "simple" or "quick" -- those are the highest-risk cases where Claude skips quality steps. Do NOT use for: skill-judge (evaluating/scoring existing skills for quality), superpowers:writing-skills (auto-triggered writing workflows).
development
voice-agents
Use when building voice AI agents, implementing speech-to-speech or pipeline (STT->LLM->TTS) architectures, optimizing voice latency, integrating voice activity detection, or designing turn-taking and barge-in handling. NEVER use for text-only chatbots, pre-recorded IVR menu trees, or music/audio processing pipelines.
development
vulnerability-scanner
Use when performing vulnerability assessments, scanning codebases for security weaknesses, mapping attack surfaces, or prioritizing security risks by exploitability. Also use when the user mentions vulnerability scanning, security audit scope, OWASP Top 10 2025, supply chain security, attack surface mapping, CVSS/EPSS scoring, or risk prioritization. NEVER use for writing secure code patterns (use security-best-practices), penetration testing execution against live systems, or compliance certification documentation (use data-privacy-compliance).
development
writing-clearly-and-concisely
Use when writing or editing any prose a human will read: documentation, commit messages, error messages, UI text, reports, README files, comments, explanations, or pull request descriptions. Use when asked to improve, tighten, or clarify existing text. NEVER for code generation, data processing, or structured output with no prose component.
development
xlsx
Use when the task involves creating, editing, analyzing, or recalculating .xlsx, .xlsm, .csv, or .tsv files -- especially financial models, structured data exports, or formula-driven spreadsheets. NEVER for purely in-memory data analysis where no file output is needed.
testing
twilio-communications
Use when building SMS messaging, voice calls, WhatsApp Business API, or phone verification features with Twilio. Also use when the user mentions Twilio, send SMS, text message, voice call IVR, phone verification, 2FA/OTP via phone, or TwiML. NEVER use for email sending (use email-systems), push notifications, non-Twilio communication APIs, or general authentication without phone verification.
development
hq-revenue-ops
Use when evaluating Sharkitect Digital deal economics, applying client tiering rules, validating pricing against margin targets, managing pipeline governance, or coordinating deal approval workflows. Covers deal scoring, pricing validation, client tier classification, and revenue-related escalation. NEVER use for general financial modeling or P&L analysis (use financial-analyst agent), market-level pricing research (use competitive-intelligence-analyst agent), or sales outreach and prospecting (use sales-researcher agent).
tools
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes
development
ui-ux-pro-max
Use when reviewing UI/UX quality, auditing accessibility, choosing color palettes, selecting font pairings, evaluating design systems, recommending chart types, applying style-specific patterns (glassmorphism, brutalism, etc.), or generating stack-specific implementation guidelines. Triggered by: UI review, UX audit, design system, accessibility check, component review, color palette, font pairing, style selection, chart recommendation, landing page design, responsive audit, dark/light mode. NEVER for building new UIs from scratch (use frontend-design), Figma-to-code translation (use figma-implement-design), applying preset themes (use theme-factory), cleaning AI-generated code (use deslop), CRO optimization (use page-cro).
development
transcribe
Use when transcribing audio/video files to text, speech-to-text from recordings, speaker diarization, labeling speakers in interviews/meetings/podcasts, or extracting text from audio. NEVER for real-time STT/TTS pipelines or voice agent implementation (use voice-ai-development). NEVER for voice agent architecture or multi-agent voice systems (use voice-agents). NEVER for audio editing, mixing, or processing (use standard audio tools). NEVER for phone system configuration or IVR (use twilio-communications).
tools
hq-knowledge-governance
Use when auditing knowledge base health, classifying documents by criticality (K1-K5), checking cross-reference integrity, identifying documentation gaps, enforcing governance compliance, or when the user asks about knowledge organization, document freshness, or KB audits. NEVER use for generic documentation writing (use documentation-templates skill), generic knowledge base design for non-Sharkitect projects (use knowledge-management skill), or research synthesis from web sources (use research-synthesizer agent).
development
hook-development
Use when user asks to create a hook, add PreToolUse/PostToolUse/Stop/SubagentStop/SessionStart/SessionEnd/UserPromptSubmit/PreCompact/Notification hook, validate tool use, implement prompt-based hooks, use ${CLAUDE_PLUGIN_ROOT}, set up event-driven automation in Claude Code, block dangerous commands via hooks, build a Claude Code plugin with hooks, or configure hooks.json. NEVER use for general webhook/API endpoint development, GitHub Actions, CI/CD pipelines, or non-Claude-Code event systems.
tools
form-cro
Use when optimizing non-signup forms for conversion: lead capture forms, contact forms, demo request forms, application forms, survey forms, quote request forms, or checkout forms. Also for form field reduction decisions, multi-step form design, form error handling optimization, or form abandonment diagnosis. NEVER for signup/registration forms (use signup-flow-cro), popups containing forms (use popup-cro), page-level conversion optimization beyond the form (use page-cro), A/B test statistical setup (use ab-test-setup).
testing
firecrawl
Use when scraping web pages, searching the web via Firecrawl CLI, crawling sites for bulk content extraction, or automating browser interactions for content behind pagination or logins. Also use when the user mentions firecrawl, web scraping to markdown, or site mapping. NEVER use for API-based data fetching (use direct HTTP), non-web content extraction, or when the user has not installed Firecrawl.
tools
feedback-mastery
Use when the user needs to prepare for a difficult conversation, give constructive feedback, navigate workplace conflict, or design a feedback system. Covers framework selection, timing decisions, power dynamics, and cultural adaptation. Do NOT use for: email composition or tone (use email-composer), internal team announcements (use internal-comms), professional writing style (use professional-communication), or meeting facilitation (use daily-meeting-update).
development
email-sequence
Use when designing a multi-email sequence, drip campaign, nurture flow, onboarding series, re-engagement campaign, or any automated email program with 3+ emails. Use when the user says 'email sequence,' 'drip campaign,' 'nurture sequence,' 'welcome series,' 'onboarding emails,' 're-engagement flow,' 'email automation,' or 'lifecycle emails.' NEVER for single one-off emails (use copywriting). NEVER for in-app onboarding flows (use onboarding-cro). NEVER for transactional emails (order confirmations, password resets) -- those are system messages, not sequences.
tools
docx
Use when creating, editing, analyzing, or reviewing .docx files, implementing tracked changes (redlining), extracting text from Word documents, or converting documents to images. NEVER use for .xlsx spreadsheets (use xlsx skill), .pptx presentations (use pptx skill), or PDF-only operations.
documentation
documentation-templates
Use when asked to write, generate, or structure documentation of any kind: READMEs, API docs, changelogs, ADRs, code comments, or llms.txt files. NEVER for writing content that documents a single function inline (use code comment conventions instead).
development
deep-interview
Use when the user's request is vague, underspecified, or ambitious. Use when the user says "I want to build...", "I need...", "Can you create...", "Let's make..." without specifying scope, constraints, success criteria, or edge cases. Use when brainstorming would start but requirements feel thin. Use when the user describes a WHAT without a WHY or WHEN or HOW MUCH. Use when you detect 2+ undefined dimensions in a request. Use when the user explicitly says "/deep-interview", "interview me", "help me think through this", "clarify requirements", "what am I missing", "poke holes in this idea". Use BEFORE brainstorming or writing-plans when the idea has gaps. Do NOT use when requirements are already specific and complete. Do NOT use for simple tasks (rename a variable, fix a typo, add a comment). Do NOT use when the user says "just do it" or provides a detailed spec. Do NOT use for debugging, code review, or operational tasks. Use superpowers:brainstorming instead when requirements are clear but design needs exploration. Use writing-plans when both requirements and design are clear.
development
session-checkpoint
Use when user wants a mid-session quick save -- 'checkpoint', 'save progress', 'back up before risky changes', '/session-checkpoint', or before compacting context. Does git commit + push + Supabase brain sync. That's it -- no audit, no lessons capture, no .tmp/ audit, no self-kill. Get back to work fast. Do NOT use for: end-of-session wrap-up (use end-session for the full 9-step audit + cleanup), skill syncing (use sync-skills.py), document freshness reviews (use document-lifecycle), resource gap detection (use resource-auditor).
testing
copywriting
Use when writing, rewriting, or improving marketing copy for ANY page type -- homepage, landing page, pricing, feature, about, or product page. Trigger on: 'write copy for,' 'improve this copy,' 'rewrite this page,' 'marketing copy,' 'headline help,' 'CTA copy,' 'page copy,' 'conversion copy,' or any request involving persuasive web page text. NEVER for email sequences (use email-sequence), popup/modal copy (use popup-cro), or line-by-line editing of existing copy (use copy-editing).
development
resource-auditor
Use after completing ANY significant deliverable, task, or output in ANY workspace. Detects four gap types: (1) UNUSED -- available skills, docs, or tools that were relevant but not invoked, (2) PROCESS -- methodology skills not followed for the task type (brainstorming skipped for creative work, systematic-debugging skipped for bugs, writing-plans skipped for multi-step work), (3) MISSING -- no purpose-built resource exists and Claude had to improvise with general knowledge, (4) FALLBACK -- generic/adjacent resource used when a specialized tool would produce meaningfully better results. Reads tool usage journal to verify methodology compliance. Writes structured gap reports to the Skill Management Hub for permanent resolution. Do NOT use for: mid-task skill discovery (check ACTIVE_SKILLS in CLAUDE.md), content enforcement in HQ (use hq-content-enforcer), code review or testing (use code-reviewer or test-engineer agents).
tools
humanizer
Remove signs of AI-generated writing from text. Use when editing or reviewing text to make it sound more natural and human-written. Based on Wikipedia's comprehensive "Signs of AI writing" guide. Detects and fixes patterns including: inflated symbolism, promotional language, superficial -ing analyses, vague attributions, em dash overuse, rule of three, AI vocabulary words, passive voice, negative parallelisms, and filler phrases.
testing
hq-content-enforcer
Use when creating, writing, rewriting, editing, reviewing, or optimizing ANY content that will be seen by clients, prospects, or the public in the HQ workspace. Includes landing pages, website copy, forms, emails, proposals, case studies, social posts, presentations, blog posts, ad copy, sales scripts, and any AI-generated text for external use. MUST be invoked BEFORE writing any content. NEVER use for internal documentation, technical specs, or code comments (no brand enforcement needed). NEVER use for content strategy planning without actual content production (use content-strategist agent).
tools
writing-plans
Use when you have a spec or requirements for a multi-step task, before touching code
development
daily-meeting-update
Use when user says 'daily', 'standup', 'scrum update', 'status update', 'what did I do yesterday', 'prepare for meeting', 'morning update', or 'team sync'. Generates formatted standup updates through interactive interview + optional tool integrations (GitHub, Jira, Claude Code history). Do NOT use for: meeting transcription or analysis (meeting-insights-analyzer), writing internal announcements (internal-comms), crafting professional messages (professional-communication), giving or receiving performance feedback (feedback-mastery).
tools
writing-plans
Use when you have a spec or requirements for a multi-step task and need a written implementation plan before touching any code. NEVER use for single-file changes or trivial edits -- inline steps are sufficient. NEVER use after implementation has already started.
development
viral-generator-builder
Build shareable generator tools (quizzes, calculators, name generators, avatar creators) optimized for viral sharing. Use when: (1) user wants to build a tool that generates personalized results, (2) user asks about viral mechanics for interactive content, (3) user needs share card or OG image design for generated results, (4) user wants quiz scoring or result distribution logic. Do NOT use for: general frontend development without viral/sharing focus (use frontend-design), marketing psychology theory without implementation (use marketing-psychology), free tool strategy without generator mechanics (use free-tool-strategy), content creation without interactive generation (use content-creator).
tools
voice-ai-development
Use when implementing voice AI features: real-time voice agents, STT/TTS pipelines, WebRTC audio, voice provider integration (OpenAI Realtime, Deepgram, ElevenLabs, Vapi, LiveKit), latency optimization, barge-in detection, or voice-specific error handling. NEVER for voice agent architecture decisions without implementation (use voice-agents), general audio processing without voice AI context (use standard libraries), phone system configuration (use twilio-communications), transcription-only workflows (use transcribe).
development
agents-crewai
Use when building multi-agent systems with CrewAI framework. Use when choosing between CrewAI Crews (autonomous) vs Flows (event-driven). Use when designing agent roles/goals/backstories for CrewAI. Use when debugging CrewAI crew failures (infinite loops, agents ignoring context, token cost explosion). Use when configuring CrewAI with different LLM providers (OpenAI, Anthropic, local). NEVER for general agent architecture decisions (use ai-agents-architect). NEVER for non-CrewAI frameworks like LangGraph or AutoGen.
development
agents-llamaindex
Use when building LLM applications with LlamaIndex: RAG pipelines, document ingestion, vector index construction, query engine configuration, or agentic retrieval. Also use when choosing between LlamaIndex index types, debugging retrieval quality, implementing advanced RAG patterns (hybrid search, reranking, routing), or selecting chunking strategies. NEVER use for LangChain-specific patterns or architecture, general prompt engineering without retrieval, non-LLM data pipelines, or vector database administration without LlamaIndex.
development
senior-backend
When designing backend system architecture, choosing between API styles, or making database scaling decisions. When reviewing backend code for production readiness, performance bottlenecks, or architectural anti-patterns. Also when planning authentication architecture or error handling strategy across service boundaries. For code quality patterns use clean-code instead. For database-specific queries use database instead. For security reviews use security-best-practices instead. For containerization use docker-expert instead. For NestJS-specific work use nestjs-expert instead. For debugging errors use error-resolver instead.
development
shopify-development
Build Shopify apps, extensions, themes using GraphQL Admin API, Shopify CLI, and Liquid. TRIGGER: "shopify", "shopify app", "checkout extension", "admin extension", "POS extension", "shopify theme", "liquid template", "polaris", "shopify graphql", "shopify webhook", "shopify billing", "app subscription", "metafields", "shopify functions" EXCLUDE: General e-commerce advice without Shopify context, payment processing (use stripe-best-practices), generic React/Node.js questions
tools
signup-flow-cro
Use when optimizing signup, registration, account creation, or trial activation flows. Also use when the user mentions signup conversions, registration friction, signup form optimization, free trial signup, reduce signup dropoff, or account creation flow. NEVER use for post-signup onboarding (use onboarding-cro), lead capture forms without account creation (use form-cro), or landing page optimization leading to signup (use page-cro).
tools
statistical-analysis
Use when conducting statistical hypothesis tests (t-test, ANOVA, chi-square, non-parametric), regression, correlation, or Bayesian analyses on research data. Also use when selecting appropriate statistical tests, checking and recovering from assumption violations, calculating effect sizes, conducting power analyses, or formatting results in APA style. NEVER use for machine learning model evaluation or hyperparameter tuning, A/B test design for product experiments (use ab-test-setup), data visualization without statistical inference, or exploratory data analysis without hypothesis testing.
development
stripe-best-practices
Guides Stripe integration decisions — API selection (Checkout Sessions vs PaymentIntents), Connect platform setup (Accounts v2, controller properties), billing/subscriptions, Treasury financial accounts, integration surfaces (Checkout, Payment Element), webhook handling, and migrating from deprecated Stripe APIs. Use when building, modifying, or reviewing any Stripe integration — including accepting payments, building marketplaces, integrating Stripe, processing payments, setting up subscriptions, handling webhooks, or creating connected accounts. Do NOT use for general payment processing concepts unrelated to Stripe, non-Stripe payment providers (PayPal, Square, Braintree), Stripe CLI setup and configuration, or Stripe Dashboard navigation.
tools
systematic-debugging
Use when encountering any bug, test failure, unexpected behavior, build failure, or integration issue -- before proposing fixes. Also use when previous fix attempts have failed, or when under time pressure where guessing is tempting. NEVER use for performance optimization without an error present, feature implementation, code review, or refactoring tasks.
development
skill-judge
MANDATORY when evaluating, auditing, scoring, or reviewing Agent Skills (SKILL.md files) for quality. MANDATORY when certifying skill reference companions under ~/.claude/skills/**/references/*.md (runs Pointer-Only Validator per SoT-Reference Discipline, universal-protocols.md). Use when comparing skills, running skill quality audits, benchmarking skill effectiveness, or deciding which skills need optimization. Use when a user says 'score this skill', 'evaluate this skill', 'audit my skills', 'how good is this skill'. CRITICAL: refuses certification of PROSE-class companions; escalates BORDERLINE classifications to AI-judge pass. Do NOT use for: ultimate-skill-creator (creating new skills), skill-creator (lightweight skill creation), superpowers:writing-skills (auto-triggered writing workflows).
development
hq-strategic-ops
Use when diagnosing architectural debt across Sharkitect's business systems, identifying structural misalignment between strategy and execution, conducting enterprise-level structural diagnosis, or evaluating whether operational problems are symptoms of deeper architectural issues. Activated by specific conditions: repeated process failures, scaling bottlenecks, tool proliferation, integration breakdowns, or when the same problem surfaces in 3+ departments. NEVER use for day-to-day operational issues (use hq-operations skill), technology-specific architecture decisions (use hq-tech-strategy skill), or competitive market analysis (use competitive-intelligence-analyst agent).
tools
end-session
Use when user says 'end session', 'wrap up', 'stop for the day', 'done for today', 'close out', 'save session', 'wrapping up', or invokes /end-session. Runs the full 9-step end-of-session protocol: resource audit, MEMORY.md update, lessons capture, plan status, pending items, workspace checklist, .tmp/ audit, git commit+push, Supabase brain sync, session brief, summary. Final step schedules a detached self-kill of the current session ONLY (3s delay) so the window closes cleanly. Other claude.exe processes (active workspaces) are NOT touched -- orphan cleanup is handled separately by Claude-Orphan-Cleanup-Hourly with proper age safeguards. Do NOT use for: mid-session quick saves (use session-checkpoint), skill syncing (use sync-skills.py), brain memory queries (use supabase-sync.py pull), document freshness reviews (use document-lifecycle), resource gap detection (use resource-auditor).
testing
n8n-node-configuration
Use when configuring n8n nodes, understanding property dependencies, determining required fields per operation, resolving invisible/missing fields, or choosing get_node detail levels. NEVER for expression syntax (use n8n-expression-syntax), MCP tool selection (use n8n-mcp-tools-expert), or Code node JS/Python (use n8n-code-javascript or n8n-code-python).
tools
seo-optimizer
Use when optimizing website content for search rankings, conducting keyword research, auditing technical SEO issues, implementing schema markup, or planning organic content strategy. NEVER use for paid search (PPC/Google Ads), social media strategy, or general copywriting that is not search-intent-driven.
development
skills/using-sharkitect-methodology
--- name: using-sharkitect-methodology description: Use when starting any conversation in a Sharkitect workspace OR before any task involving NEW pricing, positioning, proposal, strategy, plan-execution, or schema-design work — mandates invocation of Sharkitect-specific methodology skills (pricing-strategy, marketing-strategy-pmm, smb-cfo, hq-revenue-ops, executing-plans, brainstorming) under the same anti-rationalization discipline as using-superpowers. Documentation has failed 4 times across H
testing
interactive-portfolio
Use when building developer portfolios, designer portfolios, creative showcases, or personal websites meant to convert visitors into job offers or client inquiries. Also use when the user mentions portfolio architecture, case study presentation, project showcase, hiring manager psychology, or portfolio performance optimization. NEVER use for scroll animation implementation (scroll-experience), visual design direction and typography (frontend-design), portfolio copywriting and messaging (copywriting), or Figma-to-code translation (figma-implement-design).
tools
page-cro
Use when the user wants to optimize, audit, or diagnose conversion problems on any marketing page -- homepage, landing page, pricing page, feature page, or blog post. Also use for 'CRO,' 'this page isn't converting,' 'improve conversions,' or 'why isn't this page working.' NEVER for signup/registration flows (signup-flow-cro), post-signup activation (onboarding-cro), form-specific optimization (form-cro), popup/modal optimization (popup-cro), general copy rewrites (copywriting), A/B test execution (ab-test-setup), or paywall/upgrade flows (paywall-upgrade-cro).
testing
programmatic-seo
Use when building SEO-optimized pages at scale using templates and data, designing programmatic page strategies, selecting pSEO playbooks, evaluating data defensibility for page generation, or diagnosing thin content and indexation issues with template-generated pages. Also use when the user mentions programmatic SEO, template pages, directory pages, location pages, comparison pages at scale, or integration landing pages. NEVER use for auditing existing SEO technical issues (seo-optimizer), writing individual content pieces (content-research-writer), competitor product comparison strategy (competitor-alternatives), or general copywriting for landing pages (copywriting).
development
docker-expert
Use when writing or optimizing Dockerfiles, troubleshooting container builds, configuring Docker Compose, hardening container security, or reducing image sizes. Also use when the user mentions multi-stage builds, .dockerignore, container orchestration, image optimization, or Docker networking. NEVER use for Kubernetes orchestration (use kubernetes tooling), CI/CD pipeline configuration (use devops tooling), or cloud-specific container services like ECS or Cloud Run.
tools
frontend-design
Use when building web components, pages, landing pages, dashboards, HTML/CSS layouts, React/Vue/Svelte components, or any frontend interface from scratch without a Figma source. Also use when the user asks to style, beautify, or redesign an existing web UI, or requests a visually distinctive artifact (poster, card, banner). NEVER use for Figma-to-code translation (use figma-implement-design), UX critique or design system audits (use ui-ux-pro-max), applying preset color/font themes (use theme-factory), or cleaning up AI-generated code (use deslop).
development
github-actions-creator
Use when the user wants to create, debug, or optimize a GitHub Actions workflow. TRIGGER: "github actions", "CI/CD pipeline", "workflow", "deploy workflow", "CI workflow", "github action", "reusable workflow", "composite action", "workflow_dispatch", "matrix strategy", "self-hosted runner", "OIDC deployment" EXCLUDE: General git operations, GitHub API usage, non-Actions CI systems (Jenkins, CircleCI)
development