Purpose
Systematically stress-test a target's management presentation or CIM to surface inflated projections, internal inconsistencies, omitted disclosures, and selectively framed narratives — then convert those findings into a prioritized diligence request list so the deal team spends time on the risks that actually move valuation.
When to Use
Invoke this skill when the user asks to:
- Review a CIM, management presentation, or teaser for a potential acquisition
- Pressure-test a target's financial projections or growth assumptions
- Identify what's missing or being downplayed before a management meeting
- Prepare a diligence question list based on a seller's materials
- Screen an inbound deal quickly to decide whether to advance or pass
- Sanity-check a target's story against the numbers it presents
Inputs Required
Before starting, collect:
- Target overview: What the company does, sector, and approximate revenue/EBITDA size
- Materials provided: The CIM/presentation content, or a detailed summary of its sections, financials, and claims
- Deal stage: Screening / IOI-LOI / confirmatory diligence
- Deal thesis: Why the buyer is interested (platform, add-on, turnaround, growth, etc.)
- Reported financials: Historical revenue, EBITDA, margins, and the projected forecast if available
- Known adjustments: Any add-backs, pro forma adjustments, or normalizations claimed by the seller
- Areas of concern: Anything the user already wants scrutinized
If projections or adjustments are not itemized, review whatever financial figures are provided and explicitly flag the missing detail as a disclosure gap.
Instructions
Step 1: Narrative-vs-Numbers Consistency Check
Compare the story management tells against the figures they present.
- Flag claims of "recurring" or "predictable" revenue that aren't supported by contract length, retention, or backlog data
- Flag a "diversified customer base" claim where concentration data is absent or contradicts it
- Flag "market-leading" or "high-growth market" language with no cited TAM source or growth rate
- Flag qualitative momentum ("strong pipeline," "accelerating demand") that isn't reflected in the historical trend line
- Note any section where adjectives do the work that data should be doing
Step 2: Projection Realism Assessment
Test whether the forecast is defensible.
- Growth step-change: Does projected revenue growth materially exceed the historical CAGR? A hockey-stick inflection with no named driver is a red flag
- Margin expansion: Are EBITDA margins projected to expand without a credible operational reason (scale, mix, pricing, cost-out)? Flag "margins improve because revenue grows"
- Assumption stacking: Count how many favorable assumptions must all hold simultaneously — compounding optimism inflates the terminal year
- Base-year quality: Is the forecast built off a normalized base or an unusually strong (or COVID-distorted) recent year?
- Reconcile the bridge: If a revenue or EBITDA bridge is shown, verify the pieces sum and the drivers are quantified, not just labeled
For each projection flag, state the implied assumption and what evidence would validate or kill it.
Step 3: Financial Consistency & Quality-of-Earnings Signals
Scan for internal contradictions and earnings-quality issues.
- Cross-check that figures repeated across sections (summary vs. detail vs. exhibits) actually match
- Scrutinize EBITDA add-backs: separate genuinely non-recurring items from operating costs dressed as one-time ("owner add-backs," "one-time" items that recur annually, aggressive normalizations)
- Flag gaps between EBITDA and cash flow — strong EBITDA with weak cash conversion suggests working capital or capex intensity being downplayed
- Flag margins meaningfully above sector norms without explanation
- Note any change in accounting presentation, period definition, or metric definition mid-document
Step 4: Missing Disclosure Inventory
List what a buyer needs that the materials omit or gloss over.
- Customer concentration: Top 5/10 customer share and contract terms
- Revenue quality: Recurring vs. one-time split, churn/retention, backlog
- Cost structure: Fixed vs. variable, largest cost lines, wage/input inflation exposure
- Capex & maintenance: Maintenance vs. growth capex, deferred capex, asset condition
- Working capital: Normal levels, seasonality, any recent stretch of payables
- Balance sheet & obligations: Debt, leases, contingencies, litigation, regulatory exposure
- Management & key-person risk: Dependence on the owner or a few individuals
- Mark each item present, partial, or absent
Step 5: Selective Framing & Presentation Tactics
Identify how the materials may be steering perception.
- Charts that start the y-axis above zero or cherry-pick a favorable start year
- Metrics switched between periods (revenue emphasized when EBITDA is weak, or vice versa)
- Peak-year or trailing-cherry-picked comparisons
- Averages hiding a declining trend, or aggregates hiding a weak segment
- Absence of a chart or table where one would normally appear (often the most telling signal)
Step 6: Prioritized Diligence Request List
Convert every flag into an action. Produce a ranked list where each item includes:
- The concern — one sentence
- What to request — the specific document, dataset, or management question
- Why it matters — the valuation or risk implication if the concern proves true
- Priority — High (could change the decision to proceed), Medium (affects price/structure), Low (confirmatory)
Rank by potential impact on the deal decision and valuation, not by document order.
Output Format
- Structure: Six sections following the steps, with flags as labeled bullets; diligence request list as a ranked table
- Length: 900–1,300 words
- Format: Structured document — scannable bullets per section, a table for the diligence list
- Sections: Narrative-vs-Numbers → Projection Realism → Financial Consistency → Missing Disclosures → Selective Framing → Prioritized Diligence Requests
- Tone: Skeptical, precise, and commercial — the voice of a deal professional who assumes the seller's materials are a sales document, not neutral fact
Quality Checklist
Before delivering output, verify:
- [ ] Every red flag names the specific claim, figure, or omission it refers to — not a generic caution
- [ ] Projection flags state the implied assumption and the evidence that would confirm or refute it
- [ ] Add-backs are separated into genuinely non-recurring vs. questionable
- [ ] Each missing-disclosure item is marked present, partial, or absent
- [ ] The diligence list is ranked by deal impact and every item has a specific request and rationale
- [ ] At least one High-priority item is identified, or the report explicitly states no deal-breakers were found
Notes
- Treat a CIM as a marketing document: the absence of a standard exhibit (customer concentration, cash conversion, cohort/retention) is often a stronger signal than anything present
- Owner add-backs and "one-time" items that recur every year are the most common source of overstated EBITDA — scrutinize them first
- Calibrate depth to deal stage: screening should surface deal-breakers fast; confirmatory diligence should generate a granular request list
- If the user provides only a summary rather than the full materials, flag that inference is limited and prioritize requesting the primary document
How to use this skill
This file is a Claude Code skill in Anthropic's standard format — markdown with YAML frontmatter at the top.
Installing in Claude Code:
~/.claude/skills/<folder-name>/SKILL.md— auto-loads in every session anywhere on your machine (recommended for skills you'll use across multiple projects)<your-project>/.claude/skills/<folder-name>/SKILL.md— auto-loads only when working in that project (recommended for project-specific skills)
Restart Claude Code or open a new session to register.
Using in Claude.ai, ChatGPT, Gemini, or other AI tools:
- Drag this file into a new chat, or copy and paste its contents
- Send your task or request as your next message — the AI will apply the skill where it fits
For Claude Code skill details: https://docs.anthropic.com/en/docs/claude-code/skills