S&P Company Tear Sheet Generator
Build audience-specific company tear sheets from S&P Capital IQ data.
The skill restricts financial data to S&P Global, requires N/A labeling, and specifies attribution and an informational disclaimer, providing partial data-flow transparency. However, it invokes an external MCP, writes to /tmp and an output directory, and does not clearly define least privilege, sensitive-data handling, user confirmation, isolation, or rollback, so points are deducted.
The workflow, calculation checks, file verification, and missing-data handling are fairly specific, making the happy path plausible from static inspection. However, it depends on the S&P MCP, the docx skill, and an unspecified runtime; no committed tests, reproducible execution evidence, or comprehensive abnormal-input feedback is provided, so the static ceiling applies.
The audiences, four templates, inputs, trigger terms, and private-company behavior are relatively clear. The trigger scope is broad, though, and non-fit cases, service-unavailable fallbacks, Chinese-language behavior, and mainland-China reachability are not sufficiently specified, so points are deducted.
Information architecture, template references, formatting constants, page-cut order, license metadata, and repository maintenance entry points are reasonably documented. Versioning, changelog, explicit maintenance ownership, installation/dependency notes, and troubleshooting are incomplete, and external-service prerequisites are not systematically documented.
The skill defines the tear-sheet objective, audience differences, calculations, validation rules, and DOCX output format in useful detail, so it could produce a usable draft in principle. Static evidence contains no representative output or execution verification; human review remains necessary and the result depends on available paid/live S&P data, so the score stays within the static limit.
The skill supplies data-integrity rules, formulas, validation steps, file manifests, and source attribution, creating limited auditability. It provides no real CI, committed test suite, execution logs, or independent corroboration, so key paths cannot be independently reproduced from the files and the static ceiling applies.
- Confirm that the S&P Capital IQ/Kensho MCP is authorized and reachable, and review the scope of company, financial, and relationship data sent or cached.
- The skill writes to /tmp and an output directory without specifying cleanup, access controls, or recovery after failure.
- Financial figures, derived metrics, relationship analysis, and disclaimers still require professional review; the static materials do not prove successful DOCX generation.
- The broad trigger language may cause false positives, and Chinese-output quality and mainland-China reachability of core external services are unspecified.
What does this skill do, and when should you use it?
This skill uses the Kensho LLM-ready API MCP server to retrieve S&P Capital IQ data and produce professional company tear sheets, one-pagers, and overviews. It supports equity research, investment banking/M&A, corporate development, and sales/business development audiences, for both public and private companies. The workflow collects company, financial, valuation, peer, transaction, earnings, and relationship data, calculates derived metrics, and validates the arithmetic. It outputs a formatted Word document, but adoption depends on access to the required data service and runtime tooling.
The skill first identifies the company and audience, with optional comparable companies and page-length preferences; it asks for an audience when none is supplied. It creates /tmp/tear-sheet/, queries S&P Global MCP tools for company information, market data, financials, segments, valuation, consensus estimates, earnings, relationships, peers, and M&A activity, and writes each result to structured intermediate files. It then reads those files, calculates margins, growth rates, free-cash-flow conversion, capital structure, and segment mix, performs validation checks, and generates a DOCX document. The document uses Arial typography, a navy company banner, a two-column key-value header, formatted financial tables, consistent bullets, and a repeated footer disclaimer. For private companies, the workflow skips stock price, beta, consensus estimates, and trading comps when appropriate.
- An equity research analyst evaluating a public company needs a compact view of financial performance, valuation, peers, and earnings takeaways.
- An investment banker preparing for an M&A process or client pitch needs a company profile covering business information, transaction activity, and strategic commentary.
- A corporate development team screening an acquisition target needs a one-page view of segments, capital structure, ownership, relationships, and strategic fit.
- A sales or business development professional preparing for a client meeting needs company context, strategic themes, and conversation starters.
- A researcher covering a private company needs a profile that emphasizes business overview, relationships, ownership, and whatever financial information is available.
What are this skill's strengths and limitations?
- Explicitly covers four distinct audiences: equity research, IB/M&A, corporate development, and sales/business development.
- Supports both public and private companies, with defined handling for private-company data gaps.
- Persists query results to structured intermediate files and performs a dedicated calculation and validation pass before document generation.
- Provides detailed Word-formatting rules, including a two-column company header, financial tables, consistent footer text, and audience-specific synthesis.
- Requires financial information to come from S&P Global tools rather than filling gaps with model knowledge.
- Requires the S&P Capital IQ Kensho MCP service, network access, and potentially a provider subscription or API key.
- Private-company outputs may lack stock, valuation, consensus, and trading-comps sections and may contain N/A or Not disclosed fields.
- The workflow depends on shell commands, temporary directories, filesystem access, Node.js, and the docx npm package rather than being chat-only.
- The supplied material does not include the audience reference files or exact MCP query mappings, so some deployment details cannot be verified from the source provided.
- No test suite, platform test results, or sample generated documents are provided in the source material.
How do you install this skill?
This skill is part of the partner-built S&P Global plugin collection in the GitHub repository anthropics/financial-services, which is licensed under Apache License 2.0. In Cowork, open Settings → Plugins → Add plugin, paste https://github.com/anthropics/financial-services, and choose the required partner or vertical plugin from the marketplace list. In Claude Code, first run claude plugin marketplace add anthropics/financial-services, then install the relevant plugin from that marketplace. The supplied README gives installation examples for other plugins but does not specify a standalone install command for the tear-sheet skill. Access to the S&P MCP integration may require a provider subscription or API key.
How do you use this skill?
After installation, provide both a company and an audience, for example: Generate an equity research tear sheet for Microsoft. The supported audiences are Equity Research, IB / M&A, Corp Dev, and Sales / BD. You may also provide specific comparable companies and a page-length preference; if no audience is given, the skill should ask for one. The runtime must be able to reach the S&P Capital IQ Kensho MCP service and write intermediate files under /tmp/tear-sheet/. The workflow then produces a Word file named in the pattern [CompanyName]_TearSheet_[Audience]_[YYYYMMDD].docx under /mnt/user-data/outputs/. The supplied material does not include credentials, concrete query responses, or the audience reference-file contents, so available sections depend on the MCP results and deployment.