Financial Model Refresh
Refresh estimates and reassess valuation after earnings, guidance, or macro changes.
The skill describes a financial-model update workflow and does not instruct external transactions, destructive actions, or other high-risk operations; however, it lacks user confirmation, data-source and data-flow disclosure, sensitive financial-data handling, isolation, rollback, and pre/post-change audit requirements, so it is far from full marks.
The steps are broadly consistent across actuals, forecasts, valuation, and reporting; however, required inputs, missing or malformed-data handling, formula checks, failure feedback, determinism, and key-path tests are not defined. Static evidence therefore supports only a sub-10 score.
The skill identifies quarterly earnings, guidance, estimate, macro, and event-driven scenarios and provides several trigger phrases; however, non-fit boundaries, input formats, model-type limits, Chinese-language support, and environment fit are unspecified, and semantic triggering may be broad.
The document is readable and layered by workflow step, with tables, notes, and output guidance; the repository also supplies installation context and an Apache-2.0 license. The skill itself lacks dependency notes, examples, FAQs, known limitations, versioning, changelog, and explicit maintenance ownership, warranting deductions.
The workflow covers the core task from identifying changes through estimates, valuation, and summary, and its templates provide some directly usable structure; however, it contains no calculation logic, data-acquisition method, Excel-operation rules, representative output, or result validation, leaving substantial manual work and review. Static calibration therefore caps this at 7.
The skill is auditable as pinned source text, and the repository provides a pinned revision, Apache-2.0 licensing, and general plugin-validation and secret-scanning CI; however, there are no tests covering this skill's key paths, execution results, or independent corroborating evidence. Static rules therefore cap this at 5.
- Do not treat the templates as a completed model update; verify source data, GAAP versus adjusted definitions, non-recurring items, estimate-change rationale, and valuation assumptions before use.
- The skill does not define permissions, confirmation, backup, or rollback for writing to Excel or other model files; obtain human approval and retain a versioned copy before making changes.
- There are no stop conditions or diagnostic messages for malformed inputs, conflicting data, missing data, or unavailable consensus information.
- Chinese-language output and mainland-China network availability are not addressed; independently verify reachability, licensing, and compliance for any external research or market-data connection.
What it does & when to use it
This skill supports equity-research model updates when new information changes a company’s outlook. It covers reported quarterly results, management guidance, analyst estimate revisions, macro changes, and event-driven developments. The workflow plugs in new data, revises forward estimates, recalculates valuation, and summarizes the effect on estimates and price targets. If an existing model is supplied, it can produce an updated Excel model; all outputs require professional review.
Identifies the update trigger; records new revenue, gross margin, operating expense, EBITDA, EPS, segment, cash, debt, share-count, capex, and working-capital information; compares old and revised fiscal-year estimates; documents changes to growth, margin, and one-time-item assumptions; recalculates DCF, P/E, EV/EBITDA, and price-target outputs; and prepares an estimate-change summary covering rating, price target, and upside or downside to the current price.
- An equity researcher plugs quarterly reported results into a company model after an earnings release.
- An analyst revises revenue growth, margin, and EPS assumptions after management issues new guidance.
- A research team evaluates valuation effects after changes in rates, foreign exchange, or commodity prices.
- An analyst updates estimates after an acquisition, restructuring, product launch, or management change and assesses whether the thesis changed.
- An investment professional compares revised estimates with consensus and rebuilds the price-target derivation.
Pros & cons
- Covers the workflow from new-data entry through valuation impact and price-target derivation.
- Explicitly separates reported figures, forward estimates, one-time items, and GAAP versus adjusted treatment.
- Includes segment detail, balance-sheet items, cash flow, and share-count changes that can affect EPS.
- Calls for estimate-history tracking, reconciliation to reported figures, and comparison with consensus.
- The SKILL.md specifies no data provider, automated filing retrieval, or market-data connection.
- It does not define a specific Excel format, formula implementation, or validated model template.
- Valuation methods, target multiples, and rating decisions still depend on user inputs and professional judgment.
- The README states that outputs are not investment, legal, tax, or accounting advice.
How to install
In Cowork, open Settings → Plugins → Add plugin, paste https://github.com/anthropics/financial-services, and select the relevant agent or vertical plugin from the marketplace list. The skill is located at plugins/agent-plugins/earnings-reviewer/skills/model-update/. The README does not provide a standalone installation command for this skill.
How to use
In a session with the relevant plugin enabled, use a trigger such as “update model,” “plug earnings,” “refresh estimates,” “update numbers for [company],” “new guidance,” or “revise estimates.” Provide the new data and an existing model if an updated Excel file is required.