Board Pack Prep

---
name: board-pack-prep
description: >-
  Upload your draft board deck and financials, and this predicts the
  questions your board is most likely to ask (grouped by who's asking and
  why), drafts evidence-backed answers, flags the ones where your support is
  thin, and writes a first-pass opening narrative — before anyone in the room
  does. Use whenever someone is preparing for a board meeting, an investor
  update, or any leadership review and wants to know what's coming before
  they walk in. Triggers include "prep me for the board meeting", "what will
  the board ask", "review this board deck", "draft my board narrative",
  "board pack prep", "what are the directors going to push on", or uploading
  a board deck/financials and asking for a gut check before it goes out.
metadata:
  author: "Aleph - getaleph.com"
  version: "1.0.0"
---

# Board Pack Prep

## What this skill does

A board meeting is a pattern-recognition exercise for the directors in the
room and a guessing game for everyone else. This skill closes that gap: it
reads your draft deck and financials the way an experienced board prepper
would, predicts what each director is likely to push on, drafts the answers
with the evidence behind them, and writes the opening narrative — so nothing
in the room is the first time you've thought about it.

This is a **communicate** skill, not a review or a judgment call: the output
is the finished artifact you actually bring into the meeting, not a
critique of one. It also composes two other skills in this library — see
"Built from two skills" below — which is the clearest example in this
collection of what it means to layer skills instead of writing one giant one.

## When to use it

- Preparing for a board meeting, investor update, or leadership review.
- Reviewing a draft board deck before it goes out, to find the weak spots
  first.
- Drafting the CFO/exec opening narrative that frames the numbers.

## Inputs

- **Your draft board deck and financials** — the deck (PowerPoint, Google
  Slides export, or PDF) and the underlying P&L/financials it's built from.
- **Board member names**, if you have them — the more specific the
  prediction, the more useful the output.
- **Optional: prior board minutes or transcripts** — sharpens the question
  prediction by surfacing what's been asked (and pushed on) before.

## Workflow

1. **Ask who's in the room and what they've pushed on before** — per
   `references/director-patterns.md`, identify each director's likely
   background/lens (operator, former CFO/banker, investor, founder, audit
   chair) and, if prior materials are provided, their specific recurring
   concerns. Do not force a persona the materials contradict. If board
   member names aren't available and the user can't supply them, proceed
   with the role personas per `references/output-template.md` and say
   plainly that the predictions are persona-generic rather than
   director-specific.
2. **Predict each director's likely questions, grouped by who's asking** —
   apply the question taxonomy in `references/director-patterns.md`
   (growth quality, margin quality, cash/runway, execution credibility,
   data quality, unit economics, people/productivity, strategic choices).
   Specific, not generic — a good question names a number.
3. **Draft evidence-backed answers, flag the ones where support is thin** —
   for each predicted question, draft a direct answer with the supporting
   evidence from the deck/financials cited. If a question has no real
   support in the materials, say so plainly rather than manufacturing an
   answer — see `references/director-patterns.md` for the exact handling.
4. **Write the opening narrative** — using a personal voice skill only if
   the user names one, and otherwise the bundled default in
   `references/write-in-my-voice.md`, draft the opening narrative per
   `references/narrative-structure.md`: the headline, what improved, what's
   under pressure, what management is doing about it, and where the board's
   input is actually needed. Never go looking for a voice skill the user
   hasn't named — a near-match by name may be an unrelated skill entirely.
5. **Format the pack** — apply the bundled standard in
   `references/excel-formatting-guidelines.md` to any financial exhibits
   pulled into the output, so the numbers
   in the pack match house formatting standards rather than however the
   source file happened to be formatted. In the standard markdown output,
   apply only its number-format rules (parenthesized negatives, one-decimal
   percents, whole dollars); the full color-provenance standard applies
   only when the user asks for an Excel or attached exhibit.

## Built from two skills

This skill doesn't do everything itself — it calls two others:

- **`references/write-in-my-voice.md`** — so the opening narrative sounds like you, not
  like a generic template. If you name a personal voice skill, this skill
  uses it for the narrative draft; otherwise it uses the bundled default in
  this file.
- **`references/excel-formatting-guidelines.md`** — so any financial tables or exhibits
  pulled into the pack follow a consistent house number/format standard
  instead of inheriting whatever formatting the source file had.

This is the pattern worth copying when you build your own skills: narrow,
single-purpose skills compose into something bigger instead of one skill
trying to do everything itself.

## Output format

Always return the result in the structure defined in
`references/output-template.md`: director-by-director predicted questions,
the evidence-backed answers with support flagged, the risk items worth
extra prep, and the opening narrative.

## A note on scope

This skill prepares the *pack and the narrative* from a plan or a completed
period's financials. It does not verify the underlying numbers are correct
— if the deck is built on a model, that's a model-audit job; the
`model-sanity-checker` skill from the same library covers it — and it does
not replace an actual dry run with your CEO or audit committee chair before
the meeting.

# Director Patterns and Question Taxonomy

## The boardroom is a pattern-recognition room

Directors rarely ask random questions. They return to patterns: recurring
missed targets, metrics they've distrusted before, execution risks they've
seen play out at other companies, cash or covenant risk, customer
concentration, retention, capital allocation, and management credibility.
The job here is to detect those patterns from the materials provided, not to
invent generic boardroom questions that could apply to any company.

## Every director has a fingerprint

A board member's questions usually reflect their background. Use this to
personalize predictions, but only where the materials support it — do not
force a persona that contradicts what's actually in front of you.

- **Former operator** — productivity, execution, staffing, process
  discipline.
- **Former CFO or banker** — liquidity, covenants, runway, cash conversion,
  downside cases.
- **Investor / board-appointed by an investor** — valuation, growth
  durability, strategic options, capital efficiency.
- **Founder** — product-market fit, customer pain, velocity, talent.
- **Audit chair** — controls, revenue recognition, accounting judgment,
  compliance, risk.

If prior board minutes or transcripts are provided, prefer their actual
recorded concerns over a generic persona guess — a director who has twice
asked about customer concentration will ask again before they ask about
something new.

## The question taxonomy

Use these categories to generate specific, evidence-anchored questions —
never a question that could apply to any company. A good predicted question
names a number.

- **Growth quality** — Is growth sustainable? Coming from price, volume,
  mix, or one-off items? Are weak segments masked by strong ones?
- **Margin quality** — Structural or timing-driven improvement? Are costs
  classified consistently period to period?
- **Cash and runway** — Does the cash forecast reconcile to EBITDA? What's
  the downside case? Is there covenant pressure?
- **Execution credibility** — Did management deliver on prior commitments?
  Are delays explained with facts, or excuses?
- **Data quality** — Are definitions consistent across periods? Do KPIs tie
  to the financials?
- **Unit economics** — Are CAC, LTV, churn, payback credible? Are cohort
  effects visible?
- **People and productivity** — Is headcount aligned to growth? Are hiring,
  attrition, and productivity trends connected?
- **Strategic choices** — Where should capital go? What should stop? What
  needs board approval?

## The question behind the question

When a director asks about a metric, they may really be testing whether
management understands the business, whether the forecast is credible,
whether a bad trend is being hidden, or whether management is asking for
trust without evidence. Note the underlying concern alongside the question
where it's not obvious.

## Handling thin support

If a likely question has no real support in the provided materials, do not
manufacture a confident-sounding answer. Use this format instead:

> **Evidence not found in the provided materials.**
> What to prepare before the meeting: {the specific file, reconciliation,
> or owner needed}.

A flagged gap is more useful than a plausible-sounding answer that falls
apart under a follow-up question.

## What makes an answer board-ready

A strong answer has four parts: a **direct answer** (one or two sentences),
**evidence** (specific numbers or facts from the materials), the **action**
management is taking, and the **risk control** — what's being monitored,
by whom, and when. An answer that only describes activity, hides behind an
average, or can't be tied to a source in the deck is not board-ready yet.

# Opening Narrative Structure

The opening narrative is what a CFO or CEO says in the first few minutes of
a board meeting to frame everything that follows. Draft it in this
structure, every time:

1. **The headline** — one sentence. What's the single most important thing
   true about this period? Not a summary of everything — the one thing.
2. **What improved** — the real wins, backed by a number each. Two or three,
   not a list of everything that went right.
3. **What's under pressure** — named directly, not buried. The board trusts
   management more, not less, for naming the pressure before being asked
   about it.
4. **What management is doing about it** — the specific action tied to each
   pressure point named above. Not "we're monitoring it" — the actual
   decision or plan.
5. **Where the board's input is actually needed** — if there's a real ask
   (a decision, a resource, a strategic call), state it plainly. A
   narrative with no ask at the end is a status update, not a board
   conversation.

## Length

Draft two versions:

- **30-second version** — the headline plus one line each on progress and
  pressure. For the very top of the meeting.
- **3-minute version** — the full structure above, in spoken paragraphs, not
  bullet points. This is meant to be said aloud, not read as a slide.

## Tone

Calm and factual. Do not oversell the wins or catastrophize the pressure —
a board can tell when a narrative is spin, and it costs more trust than the
bad news itself would have. For the sentence-level draft once this
structure is filled in, use a personal voice skill if the user has named
one; otherwise use the bundled default in `write-in-my-voice.md`, which
ships with this skill.

# write-in-my-voice (bundled sub-skill)

This is a small, standalone voice guide that `board-pack-prep` calls when
drafting the opening narrative. It ships inside this skill so the pack
builder works out of the box, but it's written to be replaced — the whole
point of a voice skill is that it should sound like *you*, not like this
default.

## How board-pack-prep uses this file

When drafting the opening narrative (see `narrative-structure.md`), apply
the voice rules below to the sentence-level writing, after the structure
(headline, progress, pressure, action, ask) is already filled in. Voice
governs *how* it's said; the structure governs *what* gets said.

## Default voice (replace this section with your own)

Until you personalize it, use this default: direct, plain-spoken, no filler.
Short sentences next to longer ones for rhythm. Numbers where numbers
matter. No corporate hedging ("we are cautiously optimistic") — say the
actual read. Never use "leverage" as a verb, "utilize," "in today's
environment," or other filler that says nothing.

## How to personalize this

Replace the default section above with your own voice rules. A good voice
guide answers:

- **Sentence rhythm** — short and punchy, or longer and explanatory? Usually
  both, in contrast with each other.
- **What you never say** — the phrases or hedges that don't sound like you.
- **How you open** — do you lead with the number, the narrative, or a
  direct claim?
- **How you close** — do you end on the ask, a reframe, or a plain summary?
- **A few real examples** — two or three sentences you've actually said or
  written in a board or leadership context. Real examples teach a voice
  guide more than a description of the voice does.

Once personalized, every skill in your library that drafts prose (this one,
a memo, a Slack update) can call this same file, so your voice stays
consistent across every artifact Claude drafts for you — one file, reused
everywhere, instead of restating your preferences in every prompt.

# excel-formatting-guidelines (bundled sub-skill)

A small, standalone formatting standard that `board-pack-prep` applies to
any financial table or exhibit it pulls into the pack, so numbers in the
output look like they came from one disciplined model, not from whatever
formatting the source file happened to have.

## How board-pack-prep uses this file

Whenever the output includes a financial table or exhibit, reformat it per
the rules below before including it, rather than passing the source
formatting through unchanged.

## The standard

- **Font:** one consistent font and size throughout every exhibit. Mixed
  fonts across pasted tables are the fastest way for a pack to look
  assembled rather than built.
- **Cell-provenance color coding** — the standard financial-modeling
  convention:
  - **Blue text** = a hardcoded input (a number typed in, not calculated).
  - **Black text** = a formula referencing cells on the same tab.
  - **Green text** = a formula referencing another tab or file.
  This lets anyone reviewing the pack see at a glance what's an assumption
  versus what's derived, without asking.
- **Number formats:**
  - Whole dollars for anything above roughly $100K; cents only where the
    unit genuinely requires precision (per-unit economics, per-share).
  - Percentages to one decimal place, consistently, in every table.
  - Negative numbers in parentheses, not a minus sign — the standard finance
    convention, and easier to scan.
  - Right-align all numbers; left-align all labels.
- **Time orientation:** periods run left to right across columns, most
  recent on the right, unless the source material already established a
  different convention the audience expects.
- **No unexplained color.** Every color used should mean something specific
  (see cell-provenance rule above) — never color a cell for decoration.

## Why this matters in a board pack specifically

A board is grading credibility as much as content. A pack where every
number is formatted identically, and where a reviewer can tell an input
from a formula at a glance, reads as disciplined. A pack with mismatched
fonts and inconsistent decimals reads as rushed — even when the underlying
numbers are exactly right.

# Board Pack Prep — Output Format

Always produce the output in exactly this structure and order.

---

## 📋 Board Pack Prep — {Company}, {Meeting date}

**Materials reviewed:** {deck, financials, and any prior materials provided}

---

### Predicted questions, by director

Group by director (or by role/persona if names weren't given). 2-3
questions each, specific and evidence-anchored — never generic.

**{Director name or role}**
1. {specific question, naming a number where possible}
2. {specific question}
3. {specific question}

*(repeat per director)*

---

### Evidence-backed answers

| Question | Answer | Supporting evidence | Confidence |
|---|---|---|---|
| {question} | {direct answer} | {specific figure/fact + source} | {High / Medium / Low} |
| {question with no support in the materials} | **Evidence not found in the provided materials.**<br>What to prepare before the meeting: {the specific file, reconciliation, or owner needed} | — | — |

Where evidence is genuinely missing, render the row exactly like the second
example above — the two-line "Evidence not found..." text from
`references/director-patterns.md` in the Answer cell, an em dash for
Confidence — rather than a manufactured answer.

---

### 🔧 Prep before the meeting

The specific gaps found above, as an action list — the file, reconciliation,
or owner needed for each thin answer.

1. {…}
2. {…}

---

### The opening narrative

**30-second version:** {headline + one line progress + one line pressure}

**3-minute version:** {full narrative per `references/narrative-structure.md`
— headline, what improved, what's under pressure, what management is doing,
where board input is needed}

---

*Sanity check: this predicts and prepares — it doesn't replace a real dry
run with your CEO or audit committee chair before the meeting. Want help
building the rehearsal agenda?*

WHAT THIS FILE IS
=================
board-pack-prep.skill — Predicts board questions, drafts evidence-backed answers, and writes the opening narrative.

A .skill file is Anthropic's packaging format for Claude Skills. It is a
standard ZIP archive, nothing more. If you are reading this in a text
editor, you are looking at the raw bytes of that archive — this build is
intentionally uncompressed so the contents stay human-readable.

LOOK INSIDE YOURSELF
====================
Rename the file from .skill to .zip and extract it, or run:  unzip board-pack-prep.skill
Every file inside is plain-text markdown — the instructions Claude reads.
There is no executable code in this package.

WHAT'S INSIDE
=============
  SKILL.md
  references/director-patterns.md
  references/excel-formatting-guidelines.md
  references/narrative-structure.md
  references/output-template.md
  references/write-in-my-voice.md
  README.txt (this file)

HOW TO USE IT
=============
Upload this file as-is to Claude: Settings -> Skills -> Add.
Or add it in Claude Cowork or Claude Code.

From Aleph (getaleph.com), the FP&A platform.
The full guide: https://www.getaleph.com/claude-skills-library
Questions: marketing@getaleph.com