copilot-skills

Reviews a financial statement (Income Statement, Balance Sheet, or Cash Flow) and identifies values that appear materially unusual compared to historical periods, peer line items, or expected financial patterns. Use when user asks to "spot outliers", "spot anomalies", "review this financial statement", "check for unusual numbers", "analyze this income statement", "do any of these expenses look out of line", "highlight anomalies for leadership", or references an attached Income Statement, Balance Sheet, or Cash Flow file and asks for analytical review. Do NOT use for: routine spreadsheet edits, formula creation, financial modeling, budget building, or audit/compliance opinions — this skill is for analytical review only.

Compiles and sends Chris McNulty's personal weekday morning briefing — a single start-of-day email covering top news, location-aware weather, Boston sports, enterprise AI news, open threads awaiting reply, priority client updates, committed tasks, and today's calendar in Pacific Time. Use when Chris asks to "send my weekday morning briefing", "give me my weekday briefing", "run my weekday briefing", "start my workday", "what's on my plate today", "brief me for today", or "what did I miss overnight" on a workday. Only run if the requesting user is Chris McNulty; otherwise use `morning-briefing`. Do NOT use for: end-of-day wrap-ups, full inbox triage of every message, scheduling or rescheduling (use schedule-meeting), deep prep for one named meeting (use meeting-intel), or weekend catch-ups unrelated to a workday start.

Recommend the most appropriate SharePoint document library for a piece of content when the user is unsure where to upload it. Use when a user asks "where should I put this?", "which library should I upload to?", "I don't know where this belongs", "what's the right place to save this?", or similar.

Convert an Excel workbook stored in SharePoint into a self-contained, branded HTML dashboard that surfaces key data, tables, and summaries as readable web content. Use when a user asks to "turn a spreadsheet into a dashboard", "publish Excel data as a webpage", "create an HTML report from a workbook", "make our data visible without Excel", or "convert a spreadsheet to a branded page".

Onboarding installer for the Copilot Cowork marketing skills library. Run this first when a user points at the copilot-cowork-marketing folder and asks to install, set up, configure, or onboard the marketing skills. Conducts a guided interview to capture company, products, ICP, MPF (Marketing and Positioning Framework), value prop, proof points, brand voice, competitors, and channels - drafting every answer from the user's own M365 data via Work IQ and asking them only to confirm or correct. Writes the populated variables.md and the SharePoint folder scaffold the other nine skills depend on. Triggers on "install the skills", "set up the marketing skills", "onboard", "configure this library", "get started".

Converts marketing content into formatted, on-brand PowerPoint presentations (.pptx) — pitch decks, campaign recaps, QBRs, exec briefings, and event decks — using the brand colors, fonts, and logo captured in brand-assets.md. Run after copywriter or repurposing-engine when the output needs to be a polished slide deck. Triggers on "create a deck", "make a PowerPoint", "build slides", "generate a pitch deck", "format as a presentation", "make a QBR deck", "turn this into slides".

Converts marketing content drafts into formatted, on-brand Word documents (.docx) — briefs, one-pagers, case studies, reports, and executive summaries — using the brand colors, fonts, and logo captured in brand-assets.md. Run after copywriter when the output needs to be a polished, client-ready Word file. Triggers on "create a Word doc", "make a one-pager", "format this as a Word document", "generate a brief", "produce a case study document", "write a report".

Builds a 30-day editorial calendar grounded in demand signals - topic briefs with target keywords, content format, funnel stage, differentiation angle, target reader, distribution channel, and CTA. Use monthly, fed by the positioning dossier and any performance data. Triggers on "editorial calendar", "content plan", "content strategy", "what should we write about", "content briefs".

Drafts publish-ready marketing copy from a content brief - LinkedIn posts, blog intros, newsletters, email sequences, X threads, and short-form video scripts in your voice, formatted for the target platform. Use whenever a brief is ready from content-strategist; batch 3-5 at a time. Triggers on "write a LinkedIn post", "draft a blog intro", "newsletter copy", "write copy", "draft this brief".

Takes a full content package (originals plus repurposed assets) and builds a 7-day distribution schedule with channel assignments, posting times by time zone, cross-promotion sequences, engagement actions, and a 30/60/90-day recycling plan. Use weekly. Triggers on "distribution plan", "posting schedule", "when should we post", "promote this content", "content calendar scheduling".

Analyzes weekly content performance data (LinkedIn analytics, Google Analytics, Search Console, email metrics, CRM pipeline) and produces a strategic report - top performers by conversion, underperformer diagnosis, channel efficiency, pattern analysis, SEO/AEO movement, and recommendations that feed back into the editorial calendar. Use weekly. Triggers on "performance report", "what's working", "content analytics", "what to cut", "ROI of content".

Builds a positioning dossier - ICP pain map, competitor messaging audit, gap analysis, voice-of-customer signals, and recommended positioning angles. Use at setup, then quarterly or whenever a competitor shifts messaging, you launch a feature, or your ICP changes. This is the foundation every other marketing skill depends on. Triggers on "positioning", "competitive analysis", "ICP research", "messaging gaps", "competitor audit".

Takes one long-form piece (blog, newsletter, podcast or webinar recap, case study) and extracts 8-10 standalone, platform-native assets across LinkedIn, X, email, blog snippets, carousels, and short-form video scripts. Use every time you publish a long-form piece. Triggers on "repurpose this", "turn this into LinkedIn posts", "break this up", "make assets from this", "atomize content".

Optimizes your content for both traditional search (Google, Bing) and AI answer engines (Google AI Overviews, ChatGPT, Perplexity, Grok). Produces metadata, heading structure, citation-ready answer blocks, FAQ, internal linking, and content-gap analysis. Use before publishing any blog post or landing page, and quarterly on top pages. Triggers on "SEO", "AEO", "optimize for search", "metadata", "keyword mapping", "get cited by AI".

When to follow up, what to send, and when to stop on [YOUR_FIRM] outbound. Defines pre-approved templated cadence steps the Cadence skill can queue under guardrails. Use when running the Cadence loop, deciding whether to queue the next step for a lead, or interpreting a reply. Anchored on timing.md (business-hour windows), templates/ (pre-approved cadence copy), and stop-conditions.md.

Detects sends, detects replies, queues pre-approved templated cadence steps as drafts in the prospect MD file (v1 is draft-only — no auto-send), and decides when to stop. Operates on records in states draft_pending_approval, sent, awaiting_reply, cadence_step_due, replied. Built for M365 Copilot Cowork — uses SearchM365, ListMessages, and FindMeetingTimes.

Compliance and content checks that gate every outbound send in the [YOUR_FIRM] sales harness. Use whenever the Composer drafts an outbound email/Teams message or the Cadence skill prepares to queue a templated step. Enforces banned-phrase scans, suppression-list lookup, CAN-SPAM (US), GDPR (EU/UK), and send-cap checks.

Drafts a first-touch or reply message for a single prospect in the [YOUR_FIRM] firm voice. Writes the draft to the prospect MD file under `## Draft (pending approval)` AND creates the matching Outlook draft via CreateDraftMessage so the operator can review and click Send. Picks up records in state=researched or state=replied; leaves them in state=draft_pending_approval. v1 NEVER calls SendDraftMessage — the human's Send click is the approval gate.

How to research a [YOUR_FIRM] sales prospect end-to-end — pull profile, pull company, surface recent signals, score ICP fit, write a dossier. Use when given a raw lead (name + company, LinkedIn URL, or email) and you need to produce a structured dossier the Composer can write against. Invoked by the Prospector loop on every state=new prospect.

The canonical list of kill switches every [YOUR_FIRM] sales-harness agent checks at the top of every loop tick and immediately before every send. Use when running the Prospector, Composer, or Cadence skills, or any time before drafting/sending outbound — to confirm no global pause, opt-out, send-cap, or reply-rate floor is active.

How to write outbound sales messages in the [YOUR_FIRM] firm voice. Use when the Composer is drafting any outbound email, reply, or templated cadence step. Anchored on a Voice DNA artifact extracted from real best-performing messages. Includes opener/follow-up/objection frames in message-templates.md and a quarterly re-extraction procedure in voice-dna-extract.md.

How the [YOUR_FIRM] sales harness reads from and writes to Outlook via the Copilot MCP. Composer uses it to save drafts. Cadence uses it to detect sends, detect replies, and queue templated follow-ups. Use when the harness needs to draft a new email, draft a threaded reply, search sent items for a send, scan inbox for a reply, or propose meeting times.

The markdown file schema for sales prospects in the [YOUR_FIRM] outbound harness. One file per lead at PROSPECT_DIR (OneDrive). Source of truth for state, dossier, drafts, conversation, and audit log. Read and written by Prospector, Composer, and Cadence skills. Use when researching a lead, drafting outbound, processing replies, or auditing a sales prospect's history.

Researches a single prospect end-to-end and writes a dossier into their MD file. Picks up records in state=new and leaves them in state=researched or state=disqualified. Built for M365 Copilot Cowork — uses SearchM365, SharePoint/OneDrive tools, and web search.

Composes and sends a daily morning briefing email covering five industry news articles, open communications needing replies, priority client updates, top committed tasks, and today's schedule. Trigger with "send my morning briefing", "morning brief", or run on a daily schedule.

Build a structured decision log from emails, meetings, and Teams messages. Use when user asks to "log decisions", "track decisions", "what did we decide", "what we agreed", "why was this decided", "decision history", or "build a decision log". Do NOT use for action item extraction (use meeting-intel), open question tracking, or general meeting summaries.

Identify potentially outdated or unmanaged content in SharePoint by analyzing activity, ownership, and content signals, and flag items that likely require review.

Build a structured, role-based onboarding path from SharePoint content when I ask for onboarding guidance, training paths, or what a new hire should read.

Review a document against required elements in SharePoint before formal review and flag missing or unclear items.

Before committing to a new engagement, SOW, project, or ask, review actual workload and flag capacity risks across calendar, sent email, and Teams. Reads the referenced SOW or proposal if one exists (docx, pdf, xlsx) to extract hours, roles, and scope. Sweeps the next 30 days of calendar, the last two weeks of sent email for explicit commitments, and recent Teams chats for open asks. Produces a recommendation with confidence score. Use when user asks: "can I take on [project]", "do I have capacity for [SOW]", "should I commit to [engagement]", "capacity check on this SOW", "pre-commitment check", "can I lead [project]", "should I sign this SOW", "am I free to take [client]", "bandwidth check", "before I agree to this", "is this a good time to add something". Do NOT use for: routine calendar review (use calendar-management), meeting scheduling (use schedule-meeting), morning briefings (use daily-briefing), or post-commitment workload rebalancing.

Prepare a two-minute stakeholder briefing from recent email and Teams history. Reviews the last 90 days of communications with a specific person and produces a structured snapshot of commitments made, open/unresolved items, and observable communication patterns — based only on visible signals, no inferred intent. Use when user asks: "prep me for my meeting with [person]", "brief me on [person]", "stakeholder context for [person]", "what's my history with [person]", "relationship history with [person]", "get me up to speed on [person] before our meeting", "what have I committed to [person]", "what's open with [person]", "stakeholder briefing", "two-minute brief on [person]". Do NOT use for: scheduling meetings (use schedule-meeting), preparing for a specific meeting's agenda from a transcript (use meeting-intel), or drafting messages to the person (use stakeholder-comms).