Private Equity Investment DD Skill
Turns pitch decks, interview notes, financials, and data-room files into evidence-driven PE due diligence reports and IC memos ready for investment committee.
SKILL.md explicitly forbids exposing raw sensitive identifiers, tokens, bank details, includes redaction rules, no external network calls or destructive defaults, transparent data flow; but the publisher is unverified, the license is all-rights-reserved (NOASSERTION), and the copyright status of the five 'local books' is not confirmed in visible evidence, so source attribution is not fully verifiable. Deductions: opaque authorization chain, no verifiable publisher.
SKILL.md instructions are self-consistent, the 13-step workflow is clear, runnable HTML validation and sensitive-scan snippets are provided, and references/self-test.md claims test prompts exist; but nothing was executed, no actual test results, failure-feedback paths, or dependency-failure behavior are shown, and the referenced files themselves do not appear in this evidence set. Deductions: no reproduced key paths; failure handling on abnormal input is asserted, not shown.
Fit and non-fit scenarios are clearly declared (excluding listed-company screening and professional sign-off), trigger descriptions cover PE/VC/M&A cases, Chinese support is complete, output is local HTML with no overseas service dependency, so environment fit is good; but the boundary between 'quick scan vs full report' depends on user request, with limited evidence of semantic trigger precision. Deductions: boundaries declared but trigger precision lacks tested evidence.
Good layering (bilingual README, TOC, workflow, examples, maintenance notes, repo structure) and progressive disclosure via references; but the LICENSE is all-rights-reserved with NOASSERTION metadata, conflicting with normal reusable-skill expectations, no version, changelog, or named maintainer/update path, and referenced files are absent from the evidence, leaving hidden assumptions. Deductions: incomplete license/versioning/governance.
Claims to convert messy diligence materials into an evidence-driven IC decision package; the methodology (evidence matrix, red-flag tiers, QoE, valuation/terms) plausibly adds marginal value over manual templates, and HTML output should be directly usable; but no representative output sample appears in the evidence, so completeness and direct usability are unverified. Deductions: value claims unverifiable under static review; no representative outputs.
references/source-books.md is cited for methodology provenance, self-test.md claims self-test prompts, and README offers runnable validation/scan snippets; but these are author claims, the referenced files are absent from the evidence, and there is no third-party execution evidence or independent reproduction. Deductions: single evidence type, mostly self-assertion.
- The license is all-rights-reserved; redistribution, modification, and commercial use are restricted — obtain written permission before enterprise adoption.
- Copyright authorization for the five 'local diligence books' is not documented in the repository, creating compliance uncertainty around the methodology provenance.
- The publisher is unverified by the FollowSkills registry; identity and long-term maintenance are unknown, with no version number or changelog.
- No executed test evidence or representative report samples are provided — validate output on a small sample before using it in formal IC materials.
- DD conclusions must not replace formal opinions from lawyers, accountants, or other professionals; the docs declare this boundary and users should honor it.
What does this skill do, and when should you use it?
This is a Codex skill for PE/VC, M&A, strategic investment, and Pre-IPO diligence. It takes messy deal materials and produces a structured evidence system rather than a restatement of the management narrative. It mandates a basic-facts pack, an evidence matrix (claim→evidence→source→status→gap→deal implication), a red-flag register, and explicit transaction actions, with a standalone single-file HTML report as the default deliverable. The methodology synthesizes five professional Chinese DD books; the repo contains only the abstracted workflows and checklists, not the source texts. When materials are insufficient it outputs a gap-driven report and data request list instead of pretending the DD is complete.
Reads user-supplied BP/CIMs, financial statements, contracts, data-room files, interview notes, and meeting minutes; classifies deal type and decision stage, then mandatorily extracts entity, cap-table, governance, and transaction facts; decomposes the investment logic into a hypothesis tree and maps every major claim into an evidence matrix; triangulates business, financial, and industry evidence; runs valuation, return, exit, and terms analysis; grades red flags P0/P1/P2 with required deal actions; and produces a conclusion-first, table-heavy single-file HTML report or IC memo that has passed HTML parsing and sensitive-content scans.
- A PE/VC team with a pitch deck, financials, and interview notes needs a full HTML DD report for internal circulation
- A deal lead needs to compress a project into a conclusion-first IC memo before committee
- Receivables are ballooning alongside revenue and the team wants a red-flag scan with severity, evidence, and deal implications
- A semiconductor Pre-IPO deal needs focused diligence on revenue quality, customer concentration, capacity, IP, and exit feasibility
- An investor needs management interview guides and a data-room request list before a site visit
What are this skill's strengths and limitations?
- Strong evidence discipline: every claim must carry source, verification status, gap, and deal implication; management claims create hypotheses, never conclusions
- Fuses five professional DD book methodologies module by module, covering business, financial, legal, valuation, terms, and post-investment — not a single flat checklist
- Default deliverable is a standalone HTML report with conclusion-first writing, heavy tables, and copy-paste HTML validation and sensitive-content scan scripts
- Red flags are severity-graded (P0/P1/P2) and every major issue must translate into a concrete deal action: proceed, pause, reprice, condition precedent, or walk away
- Clear boundaries: identifies legal/tax/specialist risks but explicitly refuses to issue final professional opinions
- License is NOASSERTION; verify licensing before commercial use
- No automated test suite; quality relies on SKILL.md rules and self-test prompts
- Report depth depends entirely on the quality of materials you supply; with sparse inputs you only get a gap-driven report
- Explicitly unsuitable for public-company annual-report screening or secondary-market stock picks, and cannot replace legal, audit, or tax sign-off
- Installation docs target Codex Skills only; deployment steps for other platforms are undocumented in the source
How do you install this skill?
Clone the repo into the Codex skills directory: mkdir -p ~/.codex/skills && git clone https://github.com/noahnan-max/private-equity-investment-dd-skill.git ~/.codex/skills/private-equity-investment-dd. Alternatively copy it with rsync. The source docs do not document installation steps for other platforms; on Agent Skills-compatible clients, place the folder in the client's skills directory.
How do you use this skill?
Trigger it with a natural-language prompt, e.g. 'Use the private-equity-investment-dd skill to generate a full HTML investment due diligence report based on the pitch deck, interviews, financials, and contracts I provided' or 'Run a PE red-flag scan and output severity, evidence, deal implication, and next verification steps.' You must supply the deal materials to analyze; the skill then executes the 14-step operating contract and workflow defined in SKILL.md.
How does this skill compare with similar options?
The source material names no competing tools or alternatives, so no factual comparison can be made.