Quarterly Earnings Update
Rapidly produce institutional-style post-earnings reports for covered companies, centered on beats, misses, estimate revisions, and thesis impact.
The skill requires citations to earnings releases, 10-Qs, call transcripts, and consensus data, with explicit source attribution. However, it lacks user-confirmation rules, least-privilege boundaries, data-flow disclosure, sensitive-data handling, external-search authorization, and rollback guidance, so points were deducted.
The workflow, report structure, and quality checklists cover the main path reasonably well. Reliability is reduced because search, DOCX generation, and Python dependencies are external or unspecified, with no key-path tests, execution evidence, or clearly defined abnormal-input feedback; the static ceiling therefore applies.
Triggers, intended audience, and non-fit cases are relatively clear, especially the requirement that the company already be covered. Boundaries for missing inputs, non-U.S. companies, unavailable consensus data, or unavailable transcripts are unclear; Chinese-language support is not stated, and core research depends on potentially restricted overseas sites.
The skill uses a layered main file plus references, templates, examples, checklists, and an Apache-2.0 license. It lacks skill-level versioning, changelog, named maintenance responsibility, update path, and concrete dependency-installation instructions; some dependencies are only named.
The intended deliverable, length, chart count, estimate updates, and filename are specified, so the skill could theoretically produce a reviewable earnings update. Effectiveness is limited by the large manual workflow, dependence on external data and additional DOCX/modeling capabilities, and absence of a verifiable representative output.
The skill mandates document-level, dated citations for figures and tables and includes verification checklists. Evidence remains prescriptive templates and prose rather than actual outputs, test suites, or independently reproducible results; repository CI does not cover the skill's analytical key paths.
- Before execution, confirm authorization to access external sites and use company, consensus, or client materials; do not send unauthorized sensitive research data to external services.
- If the latest release, 10-Q, transcript, or pre-earnings consensus cannot be obtained, missing data should not be invented; the gap should be disclosed and escalated for materials or human review.
- DOCX, charting, and modeling capabilities are not guaranteed by this skill itself; human review should verify page count, hyperlinks, figures, estimate changes, and investment conclusions.
What it does & when to use it
This skill analyzes quarterly results for companies already under coverage and produces an 8–12-page professional equity research update of about 3,000–5,000 words. It focuses on beats or misses, key metrics, guidance changes, forward estimates, and investment-thesis impact. The format calls for 1–3 summary tables, 8–12 charts, and specific clickable citations for figures, tables, and data. The primary deliverable is a DOCX report; an XLS model update is optional.
Reads the latest earnings materials, 10-Q filing, earnings-call transcript, investor presentation, consensus estimates, and prior guidance; searches for and verifies current earnings and transcript dates; analyzes metric variances, segment performance, margins, and guidance; updates forward estimates and shows old-versus-new assumptions; creates quarterly trend, operating-metric, beat/miss, estimate-revision, and valuation charts; and produces a DOCX report with summary, results analysis, guidance, thesis assessment, valuation, estimates, and a linked sources section.
- An equity research analyst needs an update within 24–48 hours after a covered company reports quarterly results.
- A research team needs to quantify revenue or EPS variance against consensus and explain the drivers.
- An analyst needs to revise forward estimates, guidance interpretation, and the investment thesis after new results.
- A client-facing researcher needs a quarterly report focused on what changed rather than extensive company background.
Pros & cons
- Defines report length, word count, table count, chart count, and page structure.
- Covers beats or misses, segments, margins, guidance, estimates, valuation, and thesis updates.
- Requires specific dated sources and clickable links for every figure and table.
- Provides a five-phase workflow from data collection through quality control.
- It is explicitly scoped to companies already under coverage and is not suitable for initiation reports.
- It requires current web research and multiple external source types, while leaving the retrieval implementation unspecified.
- It depends on Python charting libraries and a DOCX skill; no test suite or platform validation is provided.
- The 1–2-day turnaround target depends on source and tool availability.
How to install
In Claude Cowork, open “Settings → Plugins → Add plugin,” paste https://github.com/anthropics/financial-services, and select the required vertical plugin. In Claude Code, the README provides: claude plugin marketplace add anthropics/financial-services; claude plugin install equity-research@claude-for-financial-services. The source does not document a dedicated install command for this individual skill folder.
How to use
Example triggers: Create an earnings update for Nike Q3 2024; or Analyze Apple's quarterly results. Use it only when the company is already covered. For initiation reports, flash notes, or companies without prior coverage, use a different workflow. Before analysis, search for the latest earnings, verify the dates, and collect clickable source links.
Compared to similar skills
The source explicitly distinguishes this skill from initiation reports, flash notes, and other equity research workflows. Compared with initiation coverage, it is shorter, uses fewer tables and figures, assumes existing company familiarity, and focuses on quarterly changes rather than the complete company.