What does this skill do, and when should you use it?
This skill produces a self-contained 4–5-page HTML earnings preview for a single company. It combines the latest earnings transcript, competitor analysis, valuation, consensus estimates, recent news, and sector context. Research is restricted to Kensho Grounding MCP and S&P Global MCP, with generic web access explicitly prohibited. The output includes verbatim management quotes, financial charts, peer comparisons, linked claims, and a detailed calculation appendix.
Accepts one company ticker; calls S&P Global MCP for company information, earnings calls, quarterly financials, segments, prices, competitors, market capitalization, and consensus estimates; calls Kensho Grounding MCP for earnings estimates, analyst ratings, risks, recent news, and sector intelligence; writes raw responses to intermediate files; rereads and verifies those files before calculating margins, year-over-year changes, LTM/NTM P/E, and stock returns; then generates a self-contained HTML report with charts, linked claims, source URLs, and an appendix documenting calculations.
- An equity researcher preparing a single-company preview before an upcoming earnings release.
- A portfolio manager comparing a target company with five to seven relevant public peers on valuation and stock performance.
- A research team consolidating the latest earnings call, management quotations, market expectations, and material recent news into a printable note.
- A financial professional who needs every report claim and calculation linked to an underlying source or derivation.
How do you install this skill?
- Confirm that the Kensho Grounding MCP, S&P Global MCP, and required authorizations are available before use, and assess overseas-service reachability from mainland China.
- The skill writes multiple intermediate files under /tmp/earnings-preview/ without specifying permissions, retention, cleanup, or rollback.
- The report template loads Chart.js from a CDN, creating tension with the self-contained and zero-external-access requirements; resolve this before deployment.
- The skill name, final output filename, versioning, and maintenance ownership are unclear, and the template contains unfinished placeholders; all outputs require human review.
- Shell / CLI
- Network access
- Local filesystem
- MCP Server
Kensho Grounding MCP (search)S&P Global MCP (kfinance)Browser for HTML preview
The README supports installation through Cowork by opening Settings → Plugins → Add plugin and pasting https://github.com/anthropics/financial-services, then selecting the relevant components; alternatively, upload a zip containing a plugin directory. For Claude Code, the README documents claude plugin marketplace add anthropics/financial-services followed by installation of the relevant plugin. A standalone installation command for this individual skill is not documented.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/anthropics/financial-services.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/plugins/partner-built/spglobal/skills/earnings-preview-beta" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Provide a single company ticker as the skill argument or trigger input. The workflow must use Kensho Grounding MCP search and S&P Global MCP kfinance, then writes earnings-preview-[TICKER]-YYYY-MM-DD.html. The exact trigger syntax is not fully documented in the supplied source.
What are this skill's strengths and limitations?
- Covers transcript analysis, competitor landscape, valuation, estimates, news, and sector context in one workflow.
- Requires source traceability for every report claim and preserves detailed calculation logic in an appendix.
- Includes LTM and NTM P/E, common-date stock returns, year-over-year checks, and verbatim-quote rules.
- Produces a self-contained HTML report designed for roughly four to five printed pages.
- Hard dependency on Kensho Grounding MCP and S&P Global MCP; unavailable results cannot be replaced with generic web search.
- Requires shell, filesystem, network, and MCP access.
- The source provides no test suite, sample output, or platform-specific validation evidence.
- Scope is limited to a single-company earnings preview; sibling repository skills should not be treated as part of this capability.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Single-Company Earnings Preview this page ✓ Anthropic · Official | 39 · Not recommended | ★ 40k | 1mo ago | Apache-2.0 |
| Buffett Investment Thinking System | 53 · Use with care | ★ 823 | 6mo ago | — |
| Morning Meeting Brief ✓ Anthropic · Official | 52 · Use with care | ★ 40k | 1mo ago | Apache-2.0 |
| Financial Model Refresh ✓ Anthropic · Official | 47 · Use with care | ★ 40k | 1mo ago | Apache-2.0 |
| Tech Earnings Deep Dive Analyzer | 47 · Use with care | ★ 1.1k | 6mo ago | MIT |
How did FollowSkills review this skill?
The skill restricts research to Kensho Grounding MCP and S&P Global MCP and requires source capture, raw-data persistence, and calculation traceability; no malware, covert exfiltration, or destructive default is evident. Points are deducted because there is no user-confirmation flow, sensitive-data handling, dependency-security guidance, rollback plan, or explicit permission boundary, and it mandates writing to /tmp.
The phased workflow, immediate raw-data persistence, read-back requirement, and calculation checks are fairly explicit, with a data-not-available path for some missing results. Static evidence does not establish MCP availability, tests, robust tool-failure handling, or a complete output path; the CDN dependency also conflicts somewhat with the self-contained HTML claim. Under static calibration, reliability cannot exceed 10.
The audience and core scenario are clear: a 4-5 page single-company equity-research earnings preview. However, $ARGUMENTS, non-fit boundaries, semantic trigger conditions, mainland-China reachability, and Chinese-language support are not sufficiently specified. Dependence on overseas Kensho/S&P MCP services may reduce usability for mainland-China users.
The document has phased organization, data schemas, calculation rules, a report template, and Apache-2.0 license metadata. Points are deducted for the mismatch between the skill name earnings-preview-single and the beta assessed path, missing version/changelog, unclear maintenance owner, incomplete installation/dependency notes, weak troubleshooting guidance, and unfinished template placeholders.
If the specified MCP tools, filesystem, and template work as assumed, the workflow covers company profile, financials, peers, news, valuation, and HTML generation, so the core task is plausible. Static review cannot verify correctness; substantial human review remains necessary, the final HTML filename is unspecified, and news or consensus availability is uncertain. Under static calibration, effectiveness cannot exceed 7.
Requiring MCP call references, source URLs, intermediate files, formulas, and cross-checks improves auditability; repository CI and license metadata provide limited supporting context. Points are deducted because there are no skill-specific tests or third-party execution records, Kensho provenance depends on runtime responses, and static review cannot reproduce the key paths. Under static calibration, verifiability cannot exceed 5.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →