tandem

Start Tandem onboarding — set up your AI personal assistant

Run structured brainstorming sessions with named frameworks. Triggers: "brainstorm", "brainstorm ideas", "let's brainstorm", "ideation session", "think through options", "explore possibilities".

Set up service connectors (Gmail, Calendar, Notion, Monday.com, Microsoft 365)

Generate a structured daily plan with time blocks based on tasks, calendar, and priorities. Triggers: "plan my day", "daily plan", "what should I do today", "organize my day", "schedule my day", "time block", "help me plan today", "today's plan", "morning plan".

Draft and send email replies with tone matching and context awareness. Triggers: "write a reply", "draft an email", "respond to", "email back", "send an email", "compose email", "reply to", "write back to", "help me respond".

Triage and categorize inbox emails by priority. View urgent messages, action items, FYI, and archive candidates. Triggers: "check email", "triage inbox", "what's in my inbox", "any urgent emails", "email summary", "unread messages", "morning email check", "process inbox", "catch up on email".

Track expenses with photo OCR, manual entry, or email parsing. Triggers: "add expense", "track expense", "log receipt", "expense report", "scan receipt", "monthly expenses", "tax report".

Show all capabilities, connected apps, and personalized suggestions. Triggers: "what can you do", "help", "show me options", "capabilities", "menu", "explore", "what are my tools", "show me what I have".

Capture meeting outcomes with action items, decisions, and follow-ups. Triggers: "debrief my meeting", "meeting debrief", "what happened in the meeting", "capture meeting notes", "quick debrief", "debrief standup".

Prepare for upcoming meetings with attendee context, email threads, and open tasks. Triggers: "prep me for my meeting", "meeting prep", "prepare for meeting", "prep my meetings", "prep all meetings", "get ready for my meeting".

Import ChatGPT memories into Tandem

Monday.com board management — view boards, create/update items, change statuses, sync tasks. Triggers: "monday", "board", "monday.com", "check monday", "show my tasks", "update status", "create item", "my monday tasks", "what's on my board", "sync with monday".

Generate AI images from text descriptions using Gemini 2.5 Flash Image. Triggers: "generate an image", "create a picture", "make me a photo", "AI image", "nano banana", "generate image of", "draw a [thing]".

Optimize prompts for any AI engine with anti-pattern detection. Triggers: "optimize my prompt", "improve this prompt", "prompt master", "help me write a prompt", "make this prompt better", "rewrite my prompt".

Generate structured reports, summaries, and memos as professional documents. Triggers: "write a report", "create a summary", "write up", "document this", "create a memo", "generate report", "status report", "project summary".

Deep-dive research on any topic with structured output and source citations. Triggers: "research", "find out about", "deep dive", "investigate", "look into", "what do you know about", "analyze", "compare", "study", "explore topic".

Set up brand identity for documents and presentations. Triggers: "set up my brand", "brand setup", "configure brand", "company style", "use my colors", "brand my documents", "import my brand guidelines".

Capture email writing style from samples so drafts match your tone. Triggers: "set up my email style", "learn my writing style", "match my email tone", "configure email", "my emails should sound like me", "capture my tone".

Rebuild the dashboard from stats.json and kanban data. Triggers: "sync dashboard", "update dashboard", "refresh dashboard", "/sync".

Generate diagrams, charts, and visualizations using Claude's native capabilities. Triggers: "visualize this", "draw a diagram", "create a chart", "flowchart", "mind map", "show this visually", "diagram this".

Weekly reflection — review accomplishments, carried-over tasks, and plan next week's priorities. Triggers: "weekly review", "review my week", "how was my week", "weekly summary", "end of week review", "plan next week", "weekly wrap-up", "reflect on this week", "what did I accomplish", "friday review".

[WHAT IT DOES — one sentence, imperative, third-person. Start with a verb: "Manages", "Tracks", "Generates".] Use when [SPECIFIC CONTEXT — what situation calls for this skill, what the user is trying to accomplish]. Trigger on: '[EXACT PHRASE 1]', '[EXACT PHRASE 2]', '[EXACT PHRASE 3]'. Even if the user just says '[CASUAL SHORTHAND]', use this skill. Do NOT use for [ADJACENT DOMAIN that could be confused] — use [other-skill-name] instead.

Set up and use gogcli (gog) for personal Google accounts — Gmail, Drive, Calendar, Tasks, Contacts, Docs, Sheets, and more. Trigger when: user mentions gogcli, gog, personal Gmail, personal Google account, connecting Gmail, Google Drive setup, Google Calendar connection, Google Tasks, or needs CLI access to personal Google services. Also trigger when user says "connect my Gmail", "set up Google", "access my Google Drive", or references gog commands. Trigger for @gmail.com accounts or when user wants CLI-based Google access without MCP server. Do NOT use for accessing Gmail, Drive, or Calendar after the connection is established (use email-triage, document skills, or other skills that read the connected account automatically — the connector is for setup and reconnect flows only, not ongoing usage requests), setting up a managed Google Workspace organization account with admin-level permissions or org-wide delegation (use google-workspace — this skill is for personal consumer @gmail.com accounts only), or accessing non-Google services (use the appropriate connector for other providers). Do NOT use when user wants a guided multi-connector setup covering multiple services at once (use /connect which walks through each service sequentially). google-personal is for direct CLI (gogcli/gog) setup of personal Gmail specifically.

Set up and use Google Workspace CLI (gws) for organizational Google accounts. Trigger when: user mentions Google Workspace, gws, organizational Gmail, company Google account, Google admin, Google Workspace setup, or needs to connect work Google services (Gmail, Drive, Calendar, Sheets, Docs, Chat, Tasks). Also trigger when user says "connect my work email", "set up Google for my organization", "gws setup", or references gws commands. Trigger for any @company.com or non-gmail.com Google account setup. Do NOT use for personal/consumer Google account connections via @gmail.com (use google-personal — Workspace is for managed organizational accounts with admin-level service configuration and gws CLI, not personal Gmail consumer accounts), accessing Google Workspace APIs after the integration is set up (use email-triage, document skills, and other workflow skills — they use the gws connection automatically once established; the connector is for setup and reconnect flows only), or connecting non-Google enterprise tools (use the appropriate connector for Slack, Microsoft 365, or other providers). Do NOT use when user wants a guided multi-connector setup via the /connect wizard (which offers Google Workspace as one option). google-workspace skill is for direct gws CLI setup specifically.

DEPRECATED — Use /sync instead. This skill previously updated dashboard.html directly. All dashboard updates now go through the /sync finalizer which reads ~/Tandem/stats.json and rebuilds dashboard.html from source data. Do NOT activate this skill. If triggered, redirect user to /sync.

Saves information to the correct Tandem memory file when the user explicitly asks to remember something persistently. Trigger on: "remember this", "save this", "don't forget", "note that", "keep track of", "store this for later", "add to my memory", "I want you to know that", "make a note", "log this", or when proactivity engine offers to save a preference, decision, or contact. Do NOT use for conversational reminders ("remember, the meeting is at 3pm"), reading/recalling existing memory, or deleting/updating memories (use memory-delete).

Removes or updates entries in Tandem memory files when the user asks to forget, delete, or correct stored information. Trigger on: "forget this", "delete that memory", "remove", "that's not right anymore", "that's changed", "update my memory", "correct this", "I no longer [X]", "clean up my memory", "that's outdated". Do NOT use for reading memory (use recall), adding new memories (use memory-create), task list operations, clearing conversation history, or deleting files/folders outside ~/Tandem/memory/.

Background skill that scans the last 7 days of memory/sessions/ at session start to detect repeated actions and suggest automations via the proactivity engine. NOT user-invocable. Activates silently at session start before the user's first task. Do NOT activate during /begin onboarding, when fewer than 3 session log files exist, when the user immediately starts a specific task (queue suggestion instead), or when the same pattern was already dismissed twice (permanent stop).

Background behavior engine that drives proactive suggestions after task completion, user decisions, and conversation turns. Offers memory saves, automation suggestions, and guided next steps using AskUserQuestion. Supplements Personal Preferences rules. NOT user-invocable. Activates automatically — never announced to user. Do NOT activate during /begin onboarding, during active skill flows where user is mid-step, or when user has selected "Nothing else for now" in the current turn. Do NOT suggest more than 2 actions per response.

Tracks behavioral corrections and runs self-reflection protocols — captures correction rules from user feedback, promotes patterns through HOT/COLD tiers, and runs structured reflection sessions. Use when user says 'reflect', 'what did I learn', 'self-assess', 'audit memory', 'correction:', or 'I prefer X not Y'. Even if user just says 'reflect on that', 'remember my preference', or 'capture that correction' — use this skill. Do NOT use for 'evening reflection', 'end of day', or 'what did I do today' — those trigger the productivity skill's evening reflection workflow. Do NOT use for storing facts or reference information — use the recall skill for that.

Background skill that appends session summaries to ~/Tandem/memory/sessions/YYYY-MM-DD.md at session end. NOT user-invocable. Activates when user says goodbye/done, when session is clearly winding down, or when a major task sequence completes. Do NOT log during active tasks, for trivial interactions (single question or quick lookup), for sessions with only 1-2 exchanges and no real work completed, any personal/sensitive content the user shared conversationally, or when user says "remember that" (routes to memory-create) or "add to my tasks" (routes to task-manager). Do NOT auto-trigger mid-session.

Creates new Tandem project or context folders with CLAUDE.md and MEMORY.md configured from templates. Trigger on: "create a folder", "new project", "set up a workspace for", "start a project for [client]", "make a folder for", "new workspace", "create a space for", "organize [X] in its own folder", "client folder for [name]", "project folder". Do NOT use for creating folders outside ~/Tandem/, file system operations on desktop/Dropbox/cloud storage, project creation in external tools (VSCode, Notion, Monday.com), or simple file organization not requiring a workspace.

Dashboard sync finalizer. Reads ~/Tandem/stats.json and ~/Tandem/tasks/kanban.md, recomputes all statistics, and writes the JSON block in ~/Tandem/dashboard.html. Triggers when user says "sync dashboard", "update dashboard", "refresh stats", "sync my stats", "/sync", "update my numbers", "how are my stats". Also called by session-logger at session end. Do NOT activate for opening/viewing the dashboard without sync intent (just tell user to open dashboard.html), for stats about non-Tandem tools, or for editing dashboard layout/design (that's manual HTML work).

Runs structured brainstorm and ideation sessions with named frameworks. Matches: "brainstorm ideas for", "help me brainstorm", "brainstorm this", "let's brainstorm", "ideation session", "think through options for", "structured thinking about", "SCAMPER analysis", "six thinking hats on", "pros and cons of", "help me decide between", "generate ideas for", "brainstorm ways to", "explore options for", "what are my options". Do NOT use for: prompt optimization or rewriting (use prompt-master) — e.g. "improve my prompt", "optimize this for GPT", creating visualizations or diagrams (use visualize) — e.g. "draw a flowchart", "chart this data", "create a diagram", general decision-making without ideation (just answer directly) — e.g. simple "should I do X?" with obvious answer, meeting preparation or debrief (use meeting-prep / meeting-debrief), task management or planning (use task-manager / daily-plan).

Generates AI images from text descriptions using Gemini 2.5 Flash Image API. Matches: "generate an image of", "create a picture of", "make me a photo of", "generate a photo", "AI image of", "create art of", "draw a [concrete noun]", "make an illustration of", "photorealistic image of", "artistic image of", "picture of a", "image of a", "nano banana", "make it darker", "change the background", "add more detail", "edit this image", "try variations", "generate variations". Do NOT use for: diagrams, flowcharts, charts, mind maps, org charts, architecture diagrams, data visualizations (use visualize) — e.g. "draw a flowchart", "chart this data", "org chart", "diagram this process", brainstorm or ideation (use brainstorm) — e.g. "brainstorm ideas", "explore options", prompt optimization or rewriting (use prompt-master) — e.g. "improve my prompt", "optimize this prompt".

Optimizes and engineers prompts for any AI engine with intent analysis and rewrites. Matches: "optimize this prompt", "improve my prompt", "make this prompt better", "rewrite this prompt for [engine]", "help me write a prompt", "prompt engineering", "fix my prompt", "build a prompt for", "craft a prompt", "prompt template for", "analyze my prompt", "better prompt for", "prompt optimization". Do NOT use for: brainstorm or ideation sessions (use brainstorm) — e.g. "brainstorm ideas", "explore options", "think through alternatives", creating visualizations or diagrams (use visualize) — e.g. "draw a flowchart", "visualize this", "create a diagram", general writing improvement without prompt focus (just answer directly) — e.g. "improve this email", "rewrite this paragraph", meeting preparation or debrief (use meeting-prep / meeting-debrief).

Creates visualizations using native artifacts, Mermaid diagrams, or file-based output. Matches: "visualize this", "create a diagram", "draw a flowchart", "chart this data", "show me a chart", "create a visualization", "diagram this", "map this out", "mind map", "timeline of", "compare X vs Y visually", "org chart", "architecture diagram", "show the flow", "illustrate how". Do NOT use for: AI image generation, photos, pictures of objects or scenes, photorealistic images, artistic illustrations of real things (use nano-banana) — e.g. "generate an image of", "create a picture of", "draw a cat", "photo of a sunset", "make me an illustration of a dragon", brainstorm or ideation (use brainstorm) — e.g. "brainstorm ideas", "think through options", "explore alternatives", prompt optimization or rewriting (use prompt-master) — e.g. "improve my prompt", "optimize this prompt", "rewrite for ChatGPT", meeting preparation or debrief (use meeting-prep / meeting-debrief), general writing without visual output (just answer directly).

Shows the user everything their Tandem assistant can do. Displays a dynamic, personalized capability portfolio with connection status, available commands, and suggested actions. Use this skill whenever the user says: "what can you do", "show me options", "what's available", "menu", "capabilities", "what are my tools", "show me what I have", "explore", "what commands exist", or seems lost and doesn't know what to ask for. Also trigger when the user just says "hello" or "hi" for the first time in a session and seems to need direction. Do NOT use for requests where the user has a clear task intent (route directly to the appropriate skill — email-draft for drafting, task-manager for task CRUD, email-triage for inbox review, etc.), answering general questions Claude can handle without skill discovery context, or running detailed onboarding for a specific skill (each skill has its own setup flow — explore shows what is available, setup skills handle configuration).

Triggered when user wants to create, build, plan, or review a budget. Matches: "create a budget", "help me with a budget", "build a budget", "monthly budget", "project budget", "department budget", "event budget", "financial plan", "expense tracker", "budget spreadsheet", "budget template", "how much am I spending", "plan my expenses", "cost breakdown", "budget dashboard", "financial dashboard", "interactive budget", "visualize my budget". Also trigger for any request involving spreadsheets with financial calculations, revenue projections, or cost-benefit analysis. Do NOT use for analyzing past expense data without creating a new budget (use report-writer), creating budget slides for a presentation (use presentation skill — budget-wizard can supply the numbers but slide layout belongs there), writing a cost-benefit analysis document (use report-writer), or tracking time or creating timesheets (not a budget task).

Tracks receipts and expenses — captures, categorizes, and stores expense records in ~/expenses/YYYY/MM/receipts.json with Israeli VAT compliance and multi-currency support. Trigger on: 'add expense', 'log receipt', 'track this expense', 'I spent', 'receipt', 'expense report', 'how much did I spend', 'monthly expenses', 'track spending', 'log this purchase', 'add receipt', 'expense this'. Even if the user just says 'coffee 15 NIS' or 'track that last purchase', use this skill. Do NOT use for budgeting or financial planning — use a dedicated budgeting tool. Do NOT use for invoice generation or sending — use an invoicing tool. Do NOT use for bank reconciliation or importing bank statements.

Triggered when user wants to summarize, organize, or process meeting notes. Matches: "summarize this meeting", "meeting notes", "meeting summary", "process my meeting notes", "organize meeting notes", "what was decided in the meeting", "extract action items from meeting", "write up the meeting", "meeting recap", "summarize the call", "what happened in today's meeting", "notes from the meeting", "who said what", "action items from the call", "meeting minutes". Also trigger when user pastes or shares a transcript, recording notes, or call summary. Do NOT use for summarizing written documents, articles, or research papers (use report-writer), processing email inbox or drafting email summaries directly (triage use email-triage, drafting use email-draft — meeting-notes may invoke email-draft as a sub-step but does not draft emails independently), general note-taking that is not from a meeting or call (personal journal, stream of consciousness), or interview transcripts intended as formal research documents (use report-writer).

Triggered when user wants to create a presentation, slide deck, slides, or interactive HTML slide deck. Matches: "create a presentation", "make a slide deck", "build slides", "presentation for [audience]", "pitch deck", "status update slides", "training presentation", "proposal slides", "help me with a deck", "keynote slides", "PowerPoint presentation", "PPTX", "investor deck", "sales deck", "quarterly update", "interactive HTML slide deck". Also trigger when user explicitly asks for slides or a deck, even without the word "presentation". Do NOT use for written reports, memos, or documents in prose form (use report-writer), meeting notes or call summaries (use meeting-notes), budget creation (use budget-wizard — though presentation can chain to budget-wizard for cost slides), data visualization tools or dashboards (not a slide deck), or "show me" requests that mean list or table output rather than a slide presentation.

Triggered when user wants to write, create, or generate a report or analysis document. Matches: "write a report", "create a report", "status report", "analysis report", "generate a summary", "write a proposal", "research report", "weekly report", "project report", "write up findings", "document my analysis", "prepare a report on [topic]", "executive summary", "write a memo", "create a document", "write up", "one-pager", "brief". Also trigger when user wants any structured written output that isn't an email, budget, or presentation. Do NOT use for email drafting or follow-ups (use email-draft or email-followup), budget creation (use budget-wizard), slide decks or presentations (use presentation skill), processing or summarizing meeting notes (use meeting-notes skill), or single-line quick notes where no structure or sections are needed. Do NOT use when the user wants to research a topic first before writing (use topic-research to gather information, then report-writer for the document format). Trigger report-writer only when the user has content/findings to document, or explicitly says "write a report" without needing fresh research.

Triggered when user wants to write, draft, reply to, or compose an email. Matches: "reply to [sender]", "draft an email", "write a reply", "respond to this email", "compose email to [name]", "help me write an email", "draft a follow-up", "email [name] about [topic]", "send a response". Also triggered by email-triage when user selects "Draft replies to Urgent emails". Do NOT use for non-email messaging (Slack messages, Teams replies, SMS, WhatsApp), setting follow-up reminders or tracking reply status (use email-followup), or reviewing/categorizing inbox contents (use email-triage). Do NOT use for "follow up" requests that lack explicit drafting intent (no email address, recipient, or message content) — those route to email-followup first.

Triggered when user wants to track, check on, or follow up on sent emails awaiting replies. Matches: "track this email", "follow up on [email]", "did they reply", "check for responses", "pending replies", "who hasn't responded", "remind me to follow up", "overdue emails", "waiting for reply from [name]". Also triggered by email-triage when follow-up reminders are requested, and surfaced by proactivity engine during daily briefing. Do NOT use for task or project follow-ups unrelated to a specific email thread, generic reminders not tied to a sent email, or drafting the follow-up message itself (use email-draft once tracking determines a reply is needed).

Triggered when user wants to check, triage, review, or process their email inbox. Matches: "check my email", "what's in my inbox", "triage email", "any urgent emails", "process inbox", "email summary", "unread messages", "what emails do I have", "morning email check", "catch up on email". Also triggered by daily-briefing skill when email review is part of morning routine. Do NOT use for reviewing non-email messaging (Slack DMs, Teams messages, WhatsApp, SMS), checking calendars or meetings, or drafting emails (use email-draft) or setting follow-up reminders (use email-followup) without triage context first.

Capture meeting outcomes with guided debriefing. Triggers when user says "debrief my meeting", "meeting debrief", "debrief the 2pm", "quick debrief", "debrief standup", "what came out of the meeting", "log meeting outcomes", "capture meeting decisions", "debrief my call". Do NOT use for processing transcripts or raw meeting notes (use meeting-notes), preparing for upcoming meetings (use meeting-prep), or general task creation without meeting context (use task-manager).

Prepare for upcoming meetings with attendee context, open action items, relevant email threads, and agenda suggestions. Triggers when user says "prep me for my meeting", "meeting prep", "prepare for [meeting name]", "prep all my meetings", "what should I know before my 2pm", "brief me on the client meeting", "prep me". Do NOT use for post-meeting processing (use meeting-notes or meeting-debrief), creating calendar events (use google-personal or google-workspace connector), or general task review without meeting context (use daily-plan or task-manager).

Runs the connector setup phase of Tandem onboarding (Step 5). Guides through Gmail, Calendar, Notion, Monday.com, Outlook connectors. Reads state from ~/Tandem/.onboarding-state.md. Do NOT invoke directly — only called from /begin onboarding sequence via Skill tool.

Runs the final phases of Tandem onboarding: first live task (Step 7), scheduled automations (Step 8), memory population (Step 9), and wrap-up. Reads state from ~/Tandem/.onboarding-state.md. Do NOT invoke directly — only called from /begin onboarding sequence via Skill tool.

Creates the full Tandem workspace directory structure (~/Tandem/) including all subfolder CLAUDE.md and MEMORY.md files. Reads state from ~/Tandem/.onboarding-state.md for user role and use cases. Do NOT invoke directly — only called from /begin onboarding sequence via Skill tool.

Runs the full Tandem onboarding flow during a video call with the setup team. Sets up AI personal assistant from scratch: language, preferences, connectors, workspace, first task, scheduled automations. ~45 minutes. Trigger on: "set up tandem", "start onboarding", "configure my assistant", "I'm a new user", "set up my workspace", "get started with tandem". Do NOT auto-trigger — invoked only via /begin command. Do NOT use for standalone connector setup (use /connect) or ChatGPT import (use /migrate). Do NOT trigger on "set up my workspace" alone if the user already has Tandem configured — that request routes to subfolders for project-specific folder creation. Only trigger begin when the user explicitly wants full Tandem onboarding from scratch ("set up tandem", "start onboarding", "I'm a new user").

Sets up service connectors for Tandem — Gmail, Google Calendar, Google Workspace, Notion, Monday.com, Microsoft 365 (Outlook). Guides step-by-step through each connector. Trigger on: "connect gmail", "add connector", "set up email connector", "connect monday", "connect my apps", "add integration", "link my tools", "enable connector", "connect notion", "set up google calendar". Do NOT use for general connectivity ("wifi", "internet", "connect devices"), database connections, code API integrations, or in-progress /begin onboarding.

Imports ChatGPT memories into Tandem's MEMORY.md system. Guides user through Claude's memory import process and categorizes memories into Work Context, Preferences, People & Contacts, and Session Log. Trigger on: "import from chatgpt", "migrate memories", "switch from chatgpt", "bring my chatgpt data", "import my memories", "transfer from openai", "migrate from chatgpt". Do NOT use for general data migration, database migration, file imports, or any non-ChatGPT memory source.

Monday.com board management and integration. Triggers when the user says "show my Monday tasks", "Monday board", "Monday.com", "check Monday", "update Monday", "create a task in Monday", "mark [item] as done in Monday", "move [item] to [group]", "add an update to [item]", "what's on my Monday board", "sync with Monday", "Monday status", "show board [name]", "Monday items", "create item in Monday", "update status in Monday", "my Monday tasks", or any request to read, create, update, or manage items on Monday.com boards. Also triggers when another skill (task-manager, daily-plan) offers to sync with Monday.com.

Manages GTD productivity system — captures to inbox, processes items, runs daily morning reviews, evening reflections, weekly reviews, and tracks habit streaks. All data stored in ~/productivity/ as markdown files. Trigger on: 'add to inbox', 'capture', 'daily review', 'plan my day', 'morning planning', 'process inbox', 'weekly review', 'habit streak', 'track habit', 'habits today', 'add task', 'add goal', 'evening reflection', 'end of day', 'what should I do today'. Even if the user just says 'what's on my plate' or 'what did I do today', use this skill. Do NOT use for knowledge/reference storage — use the recall skill for that. Do NOT use for calendar events or reminders managed externally. Do NOT use for meeting transcription or self-assessment reviews.

Competitor tracking and analysis. Triggers when the user says "competitor analysis", "watch competitors", "track competitors", "competitor report", "what are my competitors doing", "competitive landscape", "market analysis", "competitor update", "check on [competitor name]", "how does [X] compare to [Y] in the market", "SWOT analysis", "competitive intelligence", or any request to monitor, analyze, or compare competitors in their industry or market. Do NOT use for one-off company lookups or generic internet research about a company (use topic-research), producing daily news digests or morning briefings that mention competitors in passing (use daily-briefing), writing formal competitor analysis documents or reports (use report-writer — competitor-watch feeds data, report-writer formats it), or tracking competitors the user has not onboarded into a watchlist yet without first offering to set up tracking.

Morning summary and daily briefing. Triggers when the user says "morning briefing", "daily briefing", "what's on today", "what do I have today", "show me my day", "morning summary", "brief me", "catch me up", "what did I miss", "start my day", "good morning", or any request for a summary of today's schedule, pending tasks, and priorities. Also triggers at the start of a new day session if scheduled. Do NOT use for standalone email triage without morning briefing context (use email-triage directly), deep competitor analysis or competitive intelligence updates (use competitor-watch), full topic research on a specific subject (use topic-research), or afternoon and evening check-ins where the user wants task status rather than a morning digest (use daily-plan or task-manager).

Manages NotebookLM notebooks via the notebooklm-mcp server — create notebooks, add sources, query content, generate podcasts/audio/slides/infographics, and search across notebooks. Trigger on: 'notebook', 'notebooklm', 'research notebook', 'create podcast', 'add source', 'query notebook', 'audio overview', 'study materials', 'create from notebook'. Even if the user just says 'make a podcast from that' or 'add this to my notebook', use this skill. Do NOT use for generic web searches without NotebookLM, local file summarization without notebooks, or document management unrelated to NotebookLM notebooks. If a vendor nlm skill is installed, this skill supersedes it — remove the vendor skill if conflicts occur.

Deep-dive research on any topic. Triggers when the user says "research [topic]", "look into [topic]", "find out about [topic]", "what do you know about [topic]", "investigate [topic]", "I need information on [topic]", "tell me about [topic]", "dig into [topic]", "analyze [topic]", "explore [topic]", or any request for structured information gathering, analysis, or background research on a subject. Also triggers for "compare [X] vs [Y]", "pros and cons of [topic]", or "should I choose [X] or [Y]". Do NOT use for ongoing competitor monitoring or watchlist updates (use competitor-watch), morning briefings or daily news digests (use daily-briefing), quick one-answer factual lookups that need no multi-source synthesis, or producing polished formatted research reports (use report-writer — topic-research produces raw synthesis, report-writer formats it into a document).

Captures brand identity from user-provided assets and stores it for document skills to reference. Use this skill whenever the user says: "set up my brand", "configure brand identity", "use my company colors", "brand my documents", "my brand guidelines", "import my brand", "set up company style", "use my logo", "brand setup", "my presentations should match my company style", or when the user provides brand assets (logo, PPTX template, brand guidelines PDF, color preferences). Also trigger when document skills notice no brand-profile.md exists and the user wants branded output. Do NOT use for applying brand voice to individual content tasks after setup is complete (brand configuration is read by other skills automatically), calibrating email tone specifically (use email-style — it handles email-specific voice refinement), designing visual brand assets like logos or color palettes (brand skill captures guidelines, not design output), or auditing existing content for brand compliance without a setup/update intent.

Shows users how to customize and personalize their Tandem skills for better results. Triggers when user says "customize", "customize my skills", "personalize", "how do I edit skills", "make these mine", "tailor skills to me", "configure my preferences", "skill settings". Do NOT use for brand/voice setup (use setup/brand), email style setup (use setup/email-style), or onboarding (use onboarding/begin).

Captures the user's email writing style from sample emails so the email-draft skill can match their tone automatically. Use this skill whenever the user says: "set up my email style", "learn my writing style", "match my email tone", "configure email style", "my emails should sound like me", "train my email voice", "capture my tone", or when email-draft produces output that doesn't match the user's natural style. Do NOT use for drafting a specific email (use email-draft), setting up full brand voice or identity beyond email tone (use brand), triaging or reviewing the inbox (use email-triage), or applying the calibrated style to outgoing emails after setup is complete (email-draft reads the style profile automatically — no re-setup needed).

Daily plan generation with time blocking. Triggers when the user says "plan my day", "daily plan", "create a plan for today", "time block my day", "schedule my day", "what should I work on today", "organize my day", "help me plan today", "block my calendar", "how should I spend today", or any request to create a structured daily schedule with time-blocked tasks and priorities. Do NOT use for adding, editing, or deleting tasks without planning context (use task-manager), conducting weekly retrospectives or goal reviews (use weekly-review), morning briefings that include email triage or news digest (use daily-briefing — daily-plan handles task prioritization and time-blocking only), or requesting a habit or streak report (use productivity-os). Do NOT use when the user wants a morning overview including emails, news, or calendar digest (use daily-briefing — daily-plan is for time-blocked task scheduling only, not morning summary with email triage).

Task tracking and management via Kanban board. Tasks stored in kanban.md as structured markdown, viewable in task-manager.html browser UI. Triggers when the user says "add a task", "new task", "create a task", "I need to do [X]", "remind me to [X]", "track this task", "show my tasks", "list tasks", "what are my tasks", "what's on my plate", "complete [task]", "done with [task]", "mark [task] as done", "delete task", "edit task", "change task priority", "update task", "what's overdue", "task board", "my to-do list", "open kanban", or any request to create, view, modify, complete, or delete tracked tasks. Do NOT use for generating a prioritized daily execution plan (use daily-plan), conducting weekly task retrospectives or reviews (use weekly-review), tracking habits or streaks (use productivity-os), or interpreting tasks in the context of goals or project strategy (task-manager manages task data; planning and strategy belong to other skills). Do NOT use for requests asking how to prioritize tasks, what to work on next, or how to best use today's time (use daily-plan — task-manager manages task data; prioritization and scheduling belong to daily-plan).

Weekly reflection and planning. Triggers when the user says "weekly review", "week in review", "review my week", "how did my week go", "weekly reflection", "plan next week", "weekly summary", "end of week review", "Friday review", "what did I accomplish this week", "weekly wrap-up", "reflect on this week", or any request to review the past week's accomplishments, reflect on progress, and plan priorities for the upcoming week. Do NOT use for planning today's tasks or building today's execution schedule (use daily-plan), adding or editing individual tasks without retrospective context (use task-manager), conducting monthly or quarterly strategic reviews (weekly-review is a 7-day cadence tool — longer retrospectives are out of scope), or generating a morning briefing (use daily-briefing or daily-plan).

Interactive tutorial for the documents workspace. Walks user through creating their first document step by step. Triggered by first-visit orientation when user enters documents/ folder, or when user says "teach me documents", "how do I create documents", "document tutorial", "show me how to make a report". Do NOT use for actually creating documents, budgets, reports, or presentations (use the corresponding document skills: budget-wizard, report-writer, meeting-notes, presentation), re-running the tutorial when the user has a direct task request (even if they mention documents), or answering general questions about document formats without tutorial/learning intent.

Interactive tutorial for the email workspace. Walks user through their first email triage and draft step by step. Triggered by first-visit orientation when user enters email/ folder and accepts the tutorial offer, or when user says "teach me email", "how do I use email features", "show me email triage", "email tutorial". Do NOT use for actually checking, triaging, drafting, or following up on emails (use email-triage, email-draft, or email-followup — tutorial-email teaches these skills, it does not execute them), re-running the tutorial when the user has a direct email task (even if they mention "email skill"), or general email advice questions without tutorial walkthrough intent.

Interactive tutorial for the research workspace. Walks user through their first research task step by step. Triggered by first-visit orientation when user enters research/ folder, or when user says "teach me research", "how do I use research", "research tutorial", "show me how to research a topic". Do NOT use for actually conducting research, monitoring competitors, or generating a daily briefing (use topic-research, competitor-watch, or daily-briefing respectively — tutorial-research teaches these skills, it does not execute them), activating when the user has a specific research task ready (even if they mention "research skill"), or providing general research tips without tutorial walkthrough intent.

Interactive tutorial for the tasks workspace. Walks user through creating a task, planning their day, and using the visual kanban board. Triggered by first-visit orientation when user enters tasks/ folder, or when user says "teach me tasks", "how do I manage tasks", "task tutorial", "show me the kanban board". Do NOT use for actually planning the day, managing tasks, or conducting a weekly review (use daily-plan, task-manager, or weekly-review — tutorial-tasks teaches these skills, it does not execute them), activating when the user has a specific task management operation ready (even if they mention "task skill"), or providing general productivity advice without tutorial walkthrough intent.