Private Equity Deal Screen
Quickly decide whether inbound deal flow merits a first look under your fund’s criteria.
The skill operates on user-provided deal materials and asks for missing investment criteria, with no stated external writes or transaction execution; this reduces permission risk. However, it lacks sensitive-CIM handling, data-flow disclosure, explicit confirmation, rollback, retention controls, and clarification of the persistence implied by “save screening criteria in memory,” so points are deducted.
The workflow, input types, and intended output are mostly coherent, and incomplete or inconsistent financials are to be flagged. There are no executable scripts, tests, defined missing-field rules, calculation conventions, or diagnostic abnormal-input handling; “3-part assessment” also conflicts with the four listed output components, limiting the score.
The audience, scenarios, and trigger phrases are clear, covering CIMs, teasers, and broker materials. Non-fit boundaries, minimum inputs, behavior when criteria are unavailable, and Chinese/Mainland-China environment support are unspecified, so trigger precision and environment fit remain only partially evidenced.
The document is readable and layered, with metadata, workflow steps, a criteria table, and output requirements. The repository supplies an Apache-2.0 license, official provenance, general CI validation, and contribution guidance, but the skill lacks its own version, changelog, maintenance owner, installation/dependency notes, examples, and troubleshooting; its output description is also internally inconsistent.
It can turn deal materials into extracted facts, criteria-based screening, bull and bear cases, first-call questions, and a shareable one-page memo. But it depends on user-supplied fund criteria and defines no calculation methodology, evidence citations, missing-data treatment, or comparative benefit over manual review; professional review remains necessary, so effectiveness is limited.
The skill source is auditable, and repository workflows provide plugin validation, secret scanning, and internal-reference scanning. There are no skill-specific tests, representative outputs, or third-party execution records, so key screening behavior cannot be independently reproduced from the files and the score is capped low.
- Do not treat Pass or Hard Pass as investment advice or an autonomous decision; qualified professionals should verify source materials, financial definitions, and fund criteria.
- CIMs and broker materials may contain highly sensitive information; confirm authorization, scope, retention, and deletion controls before enabling memory or persistence.
- For missing or conflicting data, label values as unknown rather than imputing or inferring them.
- Chinese output, regional regulatory considerations, and Mainland-China network reachability are not documented; validate these before cross-language or cross-region use.
What does this skill do, and when should you use it?
This skill reviews CIMs, teasers, and broker materials to extract the facts needed for an initial deal screen. It applies the fund’s investment criteria across financial, sector, geographic, valuation, customer, and management dimensions. It produces a one-page screening memo with a verdict, bull case, bear case, and first-call questions. When the criteria are unknown, it asks the user to provide them first.
Reads a supplied CIM, teaser, or deal description and extracts the company, business description, revenue, EBITDA, margins, growth, deal type, valuation, seller motivation, management continuity, customer concentration, and obvious risks. It compares the actual deal against revenue, EBITDA, margin, growth, sector, geography, deal size, valuation multiple, customer concentration, and management-continuity criteria, recording pass/fail results. It then produces a one-page memo containing a verdict of Pass, Further Diligence, or Hard Pass, plus bull-case points, bear-case concerns, and key questions for an initial call.
- A private equity associate receives a new CIM and needs a rapid go/no-go view for initial triage.
- An investment manager gets a broker teaser and wants to identify clear mismatches with the fund mandate.
- A deal team is preparing for a first call and needs the seller motivation, management rollover, and customer risks summarized.
- An investment committee member wants a concise screen showing why a deal may fit and what remains unresolved.
What are this skill's strengths and limitations?
- Covers the full initial-screen workflow from deal-fact extraction through criteria-based assessment and memo drafting.
- Explicitly flags inconsistent or incomplete financials and requires direct treatment of obvious red flags.
- Designed for common inbound materials including CIMs, teasers, and broker materials.
- It is an initial-screening workflow, not a substitute for full diligence, valuation modeling, or professional investment judgment.
- Results depend on user-provided fund criteria; the skill supplies no default target ranges.
- The source material provides no test suite, platform test results, or standalone Claude Code installation command.
- The source does not explain how screening criteria are actually persisted for future deals.
How do you install this skill?
The skill is located at plugins/vertical-plugins/private-equity/skills/deal-screening/. The README says to install the collection in Cowork by opening Settings → Plugins → Add plugin, pasting https://github.com/anthropics/financial-services, and selecting the desired vertical from the marketplace list. The provided README does not document a standalone Claude Code installation command for the private-equity vertical; it also says any directory under plugins/ can be zipped and uploaded.
How do you use this skill?
Provide a CIM, teaser, or deal description and use a trigger such as screen this deal, review this CIM, should we look at this, triage this teaser, or deal screening. If the fund’s criteria are not already known, provide the relevant revenue, EBITDA, sector, geography, valuation, and other screening thresholds first.